Table_6_Deep_Learning_Model_Selection_DEAP.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "machine_shape": "hm"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "This is the documentation for \"M1M2: Deep-Learning-Based Real-Time Emotion Recognition from Neural Activity\" paper by\n",
        "\n",
        "Sumya Akter 1†, Rumman Ahmed Prodhan 1†, Tanmoy Sarkar Pias 2* , David Eisenberg 3*\n",
        "and Jorge Fresneda Fernandez 1\n",
        "\n",
        "1.   Martin Tuchman School of Management, New Jersey Institute of Technology, Newark, NJ 07102, USA \n",
        "2.   Department of Computer Science, Virginia Tech, Blacksburg, VA 24061, USA\n",
        "3. Department of Information Systems, Ying Wu College of Computing, New Jersey Institute of Technology, Newark, NJ 07102, USA\n",
        "\n",
        "\\* Correspondence: tanmoysarkar@vt.edu (T.S.P.); de63@njit.edu (D.E.)\n",
        "\n",
        "† These authors contributed equally to this work.\n",
        "\n",
        "Published in MDPI Sensors: https://www.mdpi.com/1424-8220/22/21/8467\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "jO1MiOhv7x66"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Table 6. Results summary of two-class classifications of valence and arousal from DEAP dataset with LSTM, BiLSTM, and CNN models using\n",
        "raw EEG data"
      ],
      "metadata": {
        "id": "HH9L5wC0-u2F"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Pyeeg has been installed**"
      ],
      "metadata": {
        "id": "VQFc1dGpCzLC"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dAEVJFOgBnY_",
        "outputId": "17f64d31-7575-46cd-c66a-f75adadbf683"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting git+https://github.com/forrestbao/pyeeg.git\n",
            "  Cloning https://github.com/forrestbao/pyeeg.git to /tmp/pip-req-build-3cu6hnz8\n",
            "  Running command git clone -q https://github.com/forrestbao/pyeeg.git /tmp/pip-req-build-3cu6hnz8\n",
            "Requirement already satisfied: numpy>=1.9.2 in /usr/local/lib/python3.7/dist-packages (from pyeeg==0.4.4) (1.21.6)\n",
            "Building wheels for collected packages: pyeeg\n",
            "  Building wheel for pyeeg (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for pyeeg: filename=pyeeg-0.4.4-py2.py3-none-any.whl size=28131 sha256=d40f8b06a009c5d0c178be2075bc94c09c63af701edf1c4e002800fa25274c72\n",
            "  Stored in directory: /tmp/pip-ephem-wheel-cache-wqgllp1n/wheels/b0/23/e4/703c908bda656959957029fa145879aa79307b2545a2ef0271\n",
            "Successfully built pyeeg\n",
            "Installing collected packages: pyeeg\n",
            "Successfully installed pyeeg-0.4.4\n"
          ]
        }
      ],
      "source": [
        "pip install git+https://github.com/forrestbao/pyeeg.git"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Here is the code for mounting google drive**"
      ],
      "metadata": {
        "id": "xr7cRJkjC7s2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eI3Cq6plC6gU",
        "outputId": "0e032ce8-1d81-4c81-f0bb-3d3d9c7d5916"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install pyeeg"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dqactiPMDO0t",
        "outputId": "702114bd-37ca-46b0-9e6f-396d288a5f72"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: pyeeg in /usr/local/lib/python3.7/dist-packages (0.4.4)\n",
            "Requirement already satisfied: numpy>=1.9.2 in /usr/local/lib/python3.7/dist-packages (from pyeeg) (1.21.6)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Imported the necessery library packages**"
      ],
      "metadata": {
        "id": "1odiyKSWDlxE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import pyeeg as pe\n",
        "import pickle as pickle\n",
        "import pandas as pd\n",
        "import math\n",
        "\n",
        "from sklearn import svm\n",
        "from sklearn.preprocessing import normalize\n",
        "\n",
        "import os\n",
        "import time\n",
        "\n",
        "import pandas as pd\n",
        "import keras.backend as K\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers.convolutional import Conv1D\n",
        "from tensorflow.keras.utils import to_categorical \n",
        "from keras.layers import Flatten\n",
        "from keras.layers import Dense\n",
        "import keras\n",
        "from keras.datasets import mnist\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout, Flatten, Conv2D, MaxPooling2D\n",
        "from keras import backend as K\n",
        "from keras.models import Model\n",
        "import timeit\n",
        "from keras.layers.core import Flatten, Dense, Dropout\n",
        "from keras.layers.convolutional import Convolution1D, MaxPooling1D, ZeroPadding1D\n",
        "import warnings\n",
        "warnings.filterwarnings('ignore')\n",
        "\n",
        "from keras.layers import Convolution1D, ZeroPadding1D, MaxPooling1D, BatchNormalization, Activation, Dropout, Flatten, Dense\n",
        "from keras.regularizers import l2\n",
        "from keras.layers import Input\n",
        "from keras.layers.merge import concatenate\n",
        "from keras.layers import Input\n",
        "from keras.layers.merge import concatenate\n",
        "from keras.utils.vis_utils import plot_model\n",
        "\n",
        "from keras.layers import Input\n",
        "from keras.layers.merge import concatenate\n",
        "from keras.utils.vis_utils import plot_model"
      ],
      "metadata": {
        "id": "9KXEoIEJDUfB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Location of dat file and location for where to save preprocessed npy files**"
      ],
      "metadata": {
        "id": "EijnrkfaI9en"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "path_to_dataset_2 = '/content/drive/Shareddrives/DEAP EEG/Dataset/physiological_recordings/data_preprocessed_python/' #Location of dat file of DEAP dataset\n",
        "path_to_dataset = '/content/drive/Shareddrives/DEAP EEG/Preprocessed_data/MDPI_Raw_Data/' #where to save preprocessed npy files"
      ],
      "metadata": {
        "id": "0lkEoCCMGGOd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pickle\n",
        "data_training = []\n",
        "label_training = []\n",
        "data_testing = []\n",
        "label_testing = []\n",
        "subjects_list = ['01', '02', '03', '04', '05', '06', '07', '08', '09', '10', '11', '12', '13', '14', '15', '16', '17', '18', '19', '20', '21', '22', '23', '24', '25', '26', '27', '28', '29', '30', '31', '32']\n",
        "for sub in subjects_list:\n",
        "    path = \"/content/drive/Shareddrives/DEAP EEG/Dataset/physiological_recordings/data_preprocessed_python/s\"+sub+\".dat\"\n",
        "    x = pickle.load(open(path, 'rb'), encoding = 'latin1')\n",
        "    sub_data = x['data']\n",
        "    sub_labels = x['labels']\n",
        "    sub_data_training = sub_data[:30, :, :]\n",
        "    sub_label_training = sub_labels[:30,:] \n",
        "    sub_data_testing = sub_data[30:, :, :]\n",
        "    sub_label_testing = sub_labels[30:, :] \n",
        "    #dataset = convert_transpose(sub_eeg)\n",
        "    data_training.extend(sub_data_training)\n",
        "    label_training.extend(sub_label_training)\n",
        "    data_testing.extend(sub_data_testing)\n",
        "    label_testing.extend(sub_label_testing)\n",
        "data_training = np.array(data_training)\n",
        "label_training = np.array(label_training)\n",
        "data_testing = np.array(data_testing)\n",
        "label_testing = np.array(label_testing)"
      ],
      "metadata": {
        "id": "5jDvExcVUm0K"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data_training.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gpxE1az7U0g2",
        "outputId": "c26818e6-cce5-4957-cf8f-d18a0cd20f48"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(960, 40, 8064)"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data_testing.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ri3hmUJ4U1vu",
        "outputId": "d819b1d4-72ee-4d22-a41f-0f937f679f54"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(320, 40, 8064)"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "scaler = StandardScaler()\n",
        "print(scaler.fit(data_training[0]))\n",
        "print(scaler.mean_)\n",
        "\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "for instance in range(data_training.shape[0]):\n",
        "  data_training[instance] = scaler.transform(data_training[instance])\n",
        "\n",
        "for instance in range(data_testing.shape[0]):\n",
        "  data_testing[instance] = scaler.transform(data_testing[instance])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QtBvsImrnKRE",
        "outputId": "8722319c-82b0-4715-c429-38fc43d9e1cd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "StandardScaler()\n",
            "[ -48.41196474  -38.02811722  -27.45047511 ... -406.35030622 -399.53196009\n",
            " -393.58519423]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "Valence_Train = np.ravel(label_training[:, [0]])\n",
        "Arousal_Train = np.ravel(label_training[:, [1]])\n",
        "Domain_Train = np.ravel(label_training[:, [2]])\n",
        "Like_Train = np.ravel(label_training[:, [3]])\n",
        "\n",
        "Valence_Test = np.ravel(label_testing[:, [0]])\n",
        "Arousal_Test = np.ravel(label_testing[:, [1]])\n",
        "Domain_Test = np.ravel(label_testing[:, [2]])\n",
        "Like_Test = np.ravel(label_testing[:, [3]])"
      ],
      "metadata": {
        "id": "Po2W1fpVVEYa"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "count_0 = 0\n",
        "count_1 = 0\n",
        "Z = np.ravel(Arousal_Train)\n",
        "for i in range(len(Z)):\n",
        "  if Z[i] >= 1 and Z[i]<=4.99:\n",
        "    Z [i] = 0\n",
        "    count_0 = count_0 + 1\n",
        "  else:\n",
        "    Z [i] = 1\n",
        "    count_1 = count_1 + 1\n",
        "print(count_0,count_1)\n",
        "\n",
        "count_0 = 0\n",
        "count_1 = 0\n",
        "L = np.ravel(Arousal_Test)\n",
        "for i in range(len(L)):\n",
        "  if L[i] >= 1 and L[i]<=4.99:\n",
        "    L[i] = 0\n",
        "    count_0 = count_0 + 1\n",
        "  else:\n",
        "    L[i] = 1\n",
        "    count_1 = count_1 + 1\n",
        "print(count_0,count_1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bTYx7DVGVIf6",
        "outputId": "13e4714a-6b1d-4af0-9c6f-11eaca7bbfd8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "437 523\n",
            "89 231\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_train = to_categorical(Z)\n",
        "y_test = to_categorical(L)"
      ],
      "metadata": {
        "id": "fWQKWeUXVX3F"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf"
      ],
      "metadata": {
        "id": "m2f_0G_sVo-E"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataset = tf.data.Dataset.from_tensor_slices((data_training, y_train))\n",
        "test_dataset= tf.data.Dataset.from_tensor_slices((data_testing, y_test))"
      ],
      "metadata": {
        "id": "2E5zKzJrVrU0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataset"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zPy1V92BV7PV",
        "outputId": "a0227d08-05a4-42d9-e78c-98a8826f6790"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<TensorSliceDataset element_spec=(TensorSpec(shape=(40, 8064), dtype=tf.float64, name=None), TensorSpec(shape=(2,), dtype=tf.float32, name=None))>"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size = 256\n",
        "num_classes = 2\n",
        "epochs = 100\n",
        "\n",
        "#input_shape=(x_train.shape[1], 1)"
      ],
      "metadata": {
        "id": "JNok-kthEy8I"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Deep Learning Algorithms"
      ],
      "metadata": {
        "id": "thhGtcUzRRzr"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## LSTM"
      ],
      "metadata": {
        "id": "DBMEwFaXifj2"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Arousal"
      ],
      "metadata": {
        "id": "7p9J1Sy5sNzD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# LSTM Examples: https://keras.io/examples/nlp/bidirectional_lstm_imdb/\n",
        "# https://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/ \n",
        "# LSTM for international airline passengers problem with time step regression framing\n",
        "import numpy\n",
        "import matplotlib.pyplot as plt\n",
        "from pandas import read_csv\n",
        "import math\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers import LSTM\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.metrics import mean_squared_error\n",
        "\n",
        "# reshape input to be [samples, time steps, features]\n",
        "#trainX = numpy.reshape(trainX, (trainX.shape[0], trainX.shape[1], 1))\n",
        "#testX = numpy.reshape(testX, (testX.shape[0], testX.shape[1], 1))\n",
        "\n",
        "# create and fit the LSTM network\n",
        "model = Sequential()\n",
        "model.add(LSTM(100, input_shape=(40, 8064), return_sequences=True))\n",
        "model.add(LSTM(100))\n",
        "model.add(Dense(20))\n",
        "model.add(Dropout(0.2))\n",
        "model.add(Dense(2))\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=[\"accuracy\"])\n",
        "model.fit(data_training, y_train, epochs=100, batch_size=100, verbose=1,validation_data=(data_testing, y_test))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GNIirqhDsPIC",
        "outputId": "0eb9697f-d64d-4b3e-fed8-6701cdab6566"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "10/10 [==============================] - 7s 279ms/step - loss: 5.4251 - accuracy: 0.5167 - val_loss: 2.5796 - val_accuracy: 0.7219\n",
            "Epoch 2/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 3.1798 - accuracy: 0.5031 - val_loss: 0.9349 - val_accuracy: 0.7312\n",
            "Epoch 3/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 2.0814 - accuracy: 0.5146 - val_loss: 1.5035 - val_accuracy: 0.3219\n",
            "Epoch 4/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 1.3061 - accuracy: 0.5240 - val_loss: 0.6289 - val_accuracy: 0.7156\n",
            "Epoch 5/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.8304 - accuracy: 0.5542 - val_loss: 0.7072 - val_accuracy: 0.4875\n",
            "Epoch 6/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 0.7339 - accuracy: 0.5312 - val_loss: 0.6609 - val_accuracy: 0.6812\n",
            "Epoch 7/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.7177 - accuracy: 0.5635 - val_loss: 0.6686 - val_accuracy: 0.6656\n",
            "Epoch 8/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6992 - accuracy: 0.5771 - val_loss: 0.6831 - val_accuracy: 0.6000\n",
            "Epoch 9/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.7089 - accuracy: 0.5740 - val_loss: 0.6740 - val_accuracy: 0.6219\n",
            "Epoch 10/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6793 - accuracy: 0.6031 - val_loss: 0.6669 - val_accuracy: 0.6344\n",
            "Epoch 11/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.6783 - accuracy: 0.6323 - val_loss: 0.6786 - val_accuracy: 0.5688\n",
            "Epoch 12/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6758 - accuracy: 0.6219 - val_loss: 0.6752 - val_accuracy: 0.6156\n",
            "Epoch 13/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6709 - accuracy: 0.6365 - val_loss: 0.6653 - val_accuracy: 0.6469\n",
            "Epoch 14/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6577 - accuracy: 0.6396 - val_loss: 0.6889 - val_accuracy: 0.5344\n",
            "Epoch 15/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.6455 - accuracy: 0.6615 - val_loss: 0.6597 - val_accuracy: 0.6719\n",
            "Epoch 16/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.6438 - accuracy: 0.6750 - val_loss: 0.6855 - val_accuracy: 0.5500\n",
            "Epoch 17/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.6414 - accuracy: 0.6635 - val_loss: 0.6685 - val_accuracy: 0.6469\n",
            "Epoch 18/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 0.6335 - accuracy: 0.6687 - val_loss: 0.6903 - val_accuracy: 0.5844\n",
            "Epoch 19/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.6189 - accuracy: 0.7167 - val_loss: 0.7001 - val_accuracy: 0.5719\n",
            "Epoch 20/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.6229 - accuracy: 0.7052 - val_loss: 0.6930 - val_accuracy: 0.5688\n",
            "Epoch 21/100\n",
            "10/10 [==============================] - 1s 104ms/step - loss: 0.6057 - accuracy: 0.7104 - val_loss: 0.6968 - val_accuracy: 0.6031\n",
            "Epoch 22/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5938 - accuracy: 0.7073 - val_loss: 0.6927 - val_accuracy: 0.5656\n",
            "Epoch 23/100\n",
            "10/10 [==============================] - 1s 104ms/step - loss: 0.5894 - accuracy: 0.7115 - val_loss: 0.7267 - val_accuracy: 0.5813\n",
            "Epoch 24/100\n",
            "10/10 [==============================] - 1s 104ms/step - loss: 0.5983 - accuracy: 0.7031 - val_loss: 0.7201 - val_accuracy: 0.5594\n",
            "Epoch 25/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 0.5952 - accuracy: 0.7052 - val_loss: 0.7026 - val_accuracy: 0.5437\n",
            "Epoch 26/100\n",
            "10/10 [==============================] - 2s 175ms/step - loss: 0.5902 - accuracy: 0.7292 - val_loss: 0.6889 - val_accuracy: 0.5813\n",
            "Epoch 27/100\n",
            "10/10 [==============================] - 2s 155ms/step - loss: 0.5706 - accuracy: 0.7271 - val_loss: 0.7467 - val_accuracy: 0.4906\n",
            "Epoch 28/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5653 - accuracy: 0.7135 - val_loss: 0.7640 - val_accuracy: 0.5406\n",
            "Epoch 29/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 0.5446 - accuracy: 0.7385 - val_loss: 0.7992 - val_accuracy: 0.5125\n",
            "Epoch 30/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5449 - accuracy: 0.7250 - val_loss: 0.8332 - val_accuracy: 0.5063\n",
            "Epoch 31/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5319 - accuracy: 0.7573 - val_loss: 0.8629 - val_accuracy: 0.5031\n",
            "Epoch 32/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5233 - accuracy: 0.7552 - val_loss: 0.8838 - val_accuracy: 0.5906\n",
            "Epoch 33/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5597 - accuracy: 0.7281 - val_loss: 0.9689 - val_accuracy: 0.5938\n",
            "Epoch 34/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.5093 - accuracy: 0.7542 - val_loss: 1.0320 - val_accuracy: 0.4938\n",
            "Epoch 35/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5696 - accuracy: 0.7375 - val_loss: 0.9149 - val_accuracy: 0.6219\n",
            "Epoch 36/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6163 - accuracy: 0.7177 - val_loss: 1.1519 - val_accuracy: 0.4656\n",
            "Epoch 37/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6923 - accuracy: 0.6844 - val_loss: 0.8341 - val_accuracy: 0.5719\n",
            "Epoch 38/100\n",
            "10/10 [==============================] - 1s 104ms/step - loss: 0.6361 - accuracy: 0.7125 - val_loss: 0.9718 - val_accuracy: 0.6656\n",
            "Epoch 39/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.6552 - accuracy: 0.6812 - val_loss: 0.7986 - val_accuracy: 0.6375\n",
            "Epoch 40/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.6038 - accuracy: 0.6948 - val_loss: 0.7268 - val_accuracy: 0.5781\n",
            "Epoch 41/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6261 - accuracy: 0.6771 - val_loss: 0.7990 - val_accuracy: 0.6125\n",
            "Epoch 42/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.6571 - accuracy: 0.6865 - val_loss: 0.6454 - val_accuracy: 0.6344\n",
            "Epoch 43/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.6028 - accuracy: 0.6969 - val_loss: 0.6856 - val_accuracy: 0.5656\n",
            "Epoch 44/100\n",
            "10/10 [==============================] - 1s 110ms/step - loss: 0.5687 - accuracy: 0.7281 - val_loss: 0.7869 - val_accuracy: 0.5469\n",
            "Epoch 45/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.5526 - accuracy: 0.7333 - val_loss: 0.7602 - val_accuracy: 0.5750\n",
            "Epoch 46/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5258 - accuracy: 0.7375 - val_loss: 0.7342 - val_accuracy: 0.5844\n",
            "Epoch 47/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.4956 - accuracy: 0.7510 - val_loss: 0.8425 - val_accuracy: 0.4906\n",
            "Epoch 48/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 0.4973 - accuracy: 0.7594 - val_loss: 0.9759 - val_accuracy: 0.4094\n",
            "Epoch 49/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5065 - accuracy: 0.7490 - val_loss: 0.9040 - val_accuracy: 0.4219\n",
            "Epoch 50/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5213 - accuracy: 0.7427 - val_loss: 0.7852 - val_accuracy: 0.5781\n",
            "Epoch 51/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5605 - accuracy: 0.7260 - val_loss: 0.9180 - val_accuracy: 0.6062\n",
            "Epoch 52/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5779 - accuracy: 0.7490 - val_loss: 0.8309 - val_accuracy: 0.4625\n",
            "Epoch 53/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5182 - accuracy: 0.7677 - val_loss: 0.8799 - val_accuracy: 0.4844\n",
            "Epoch 54/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.5347 - accuracy: 0.7583 - val_loss: 0.8823 - val_accuracy: 0.6187\n",
            "Epoch 55/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.4919 - accuracy: 0.7896 - val_loss: 1.1452 - val_accuracy: 0.4594\n",
            "Epoch 56/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.4985 - accuracy: 0.7677 - val_loss: 0.9137 - val_accuracy: 0.5625\n",
            "Epoch 57/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5003 - accuracy: 0.7688 - val_loss: 0.9242 - val_accuracy: 0.5875\n",
            "Epoch 58/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 0.5260 - accuracy: 0.7844 - val_loss: 2.4542 - val_accuracy: 0.6969\n",
            "Epoch 59/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 3.4549 - accuracy: 0.6021 - val_loss: 4.3232 - val_accuracy: 0.3438\n",
            "Epoch 60/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 1.5484 - accuracy: 0.5927 - val_loss: 1.4722 - val_accuracy: 0.4375\n",
            "Epoch 61/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.9240 - accuracy: 0.5552 - val_loss: 0.6823 - val_accuracy: 0.5938\n",
            "Epoch 62/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.6944 - accuracy: 0.6083 - val_loss: 0.6510 - val_accuracy: 0.6375\n",
            "Epoch 63/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.6788 - accuracy: 0.6260 - val_loss: 0.6935 - val_accuracy: 0.5312\n",
            "Epoch 64/100\n",
            "10/10 [==============================] - 1s 109ms/step - loss: 0.6444 - accuracy: 0.6687 - val_loss: 0.6800 - val_accuracy: 0.5813\n",
            "Epoch 65/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 0.6258 - accuracy: 0.6698 - val_loss: 0.6778 - val_accuracy: 0.5875\n",
            "Epoch 66/100\n",
            "10/10 [==============================] - 1s 109ms/step - loss: 0.6127 - accuracy: 0.6771 - val_loss: 0.6718 - val_accuracy: 0.5750\n",
            "Epoch 67/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5985 - accuracy: 0.6979 - val_loss: 0.6837 - val_accuracy: 0.5656\n",
            "Epoch 68/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5896 - accuracy: 0.6927 - val_loss: 0.6971 - val_accuracy: 0.5188\n",
            "Epoch 69/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5802 - accuracy: 0.7115 - val_loss: 0.7032 - val_accuracy: 0.5250\n",
            "Epoch 70/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5703 - accuracy: 0.7094 - val_loss: 0.6949 - val_accuracy: 0.5469\n",
            "Epoch 71/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5679 - accuracy: 0.7229 - val_loss: 0.7206 - val_accuracy: 0.5156\n",
            "Epoch 72/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5505 - accuracy: 0.7146 - val_loss: 0.7471 - val_accuracy: 0.5750\n",
            "Epoch 73/100\n",
            "10/10 [==============================] - 1s 103ms/step - loss: 0.5501 - accuracy: 0.7375 - val_loss: 0.7478 - val_accuracy: 0.4750\n",
            "Epoch 74/100\n",
            "10/10 [==============================] - 1s 122ms/step - loss: 0.5334 - accuracy: 0.7427 - val_loss: 0.7559 - val_accuracy: 0.4812\n",
            "Epoch 75/100\n",
            "10/10 [==============================] - 2s 153ms/step - loss: 0.5151 - accuracy: 0.7604 - val_loss: 0.8484 - val_accuracy: 0.4875\n",
            "Epoch 76/100\n",
            "10/10 [==============================] - 2s 157ms/step - loss: 0.5068 - accuracy: 0.7604 - val_loss: 0.9048 - val_accuracy: 0.4938\n",
            "Epoch 77/100\n",
            "10/10 [==============================] - 1s 153ms/step - loss: 0.4980 - accuracy: 0.7479 - val_loss: 0.9046 - val_accuracy: 0.4844\n",
            "Epoch 78/100\n",
            "10/10 [==============================] - 1s 145ms/step - loss: 0.4855 - accuracy: 0.7552 - val_loss: 0.8627 - val_accuracy: 0.5531\n",
            "Epoch 79/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.4739 - accuracy: 0.7708 - val_loss: 1.0351 - val_accuracy: 0.4594\n",
            "Epoch 80/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.5064 - accuracy: 0.7615 - val_loss: 1.0201 - val_accuracy: 0.5188\n",
            "Epoch 81/100\n",
            "10/10 [==============================] - 1s 109ms/step - loss: 0.5181 - accuracy: 0.7583 - val_loss: 1.0736 - val_accuracy: 0.4750\n",
            "Epoch 82/100\n",
            "10/10 [==============================] - 1s 104ms/step - loss: 0.5357 - accuracy: 0.7437 - val_loss: 0.9171 - val_accuracy: 0.5531\n",
            "Epoch 83/100\n",
            "10/10 [==============================] - 1s 107ms/step - loss: 0.5567 - accuracy: 0.7354 - val_loss: 1.0212 - val_accuracy: 0.5688\n",
            "Epoch 84/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5344 - accuracy: 0.7437 - val_loss: 1.0005 - val_accuracy: 0.6406\n",
            "Epoch 85/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5015 - accuracy: 0.7469 - val_loss: 1.0180 - val_accuracy: 0.4719\n",
            "Epoch 86/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.4732 - accuracy: 0.7531 - val_loss: 1.0233 - val_accuracy: 0.5156\n",
            "Epoch 87/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.5528 - accuracy: 0.7490 - val_loss: 1.0722 - val_accuracy: 0.5063\n",
            "Epoch 88/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5732 - accuracy: 0.7531 - val_loss: 1.0773 - val_accuracy: 0.5625\n",
            "Epoch 89/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.4924 - accuracy: 0.7802 - val_loss: 1.1698 - val_accuracy: 0.5094\n",
            "Epoch 90/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 1.1484 - accuracy: 0.6948 - val_loss: 1.6849 - val_accuracy: 0.4844\n",
            "Epoch 91/100\n",
            "10/10 [==============================] - 1s 125ms/step - loss: 0.8230 - accuracy: 0.6917 - val_loss: 1.6124 - val_accuracy: 0.4375\n",
            "Epoch 92/100\n",
            "10/10 [==============================] - 1s 128ms/step - loss: 0.6493 - accuracy: 0.7115 - val_loss: 0.8850 - val_accuracy: 0.5656\n",
            "Epoch 93/100\n",
            "10/10 [==============================] - 1s 131ms/step - loss: 0.5707 - accuracy: 0.7333 - val_loss: 0.7688 - val_accuracy: 0.5375\n",
            "Epoch 94/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.5285 - accuracy: 0.7500 - val_loss: 0.7725 - val_accuracy: 0.5844\n",
            "Epoch 95/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.5044 - accuracy: 0.7781 - val_loss: 0.9254 - val_accuracy: 0.5625\n",
            "Epoch 96/100\n",
            "10/10 [==============================] - 1s 105ms/step - loss: 0.4713 - accuracy: 0.7781 - val_loss: 1.0459 - val_accuracy: 0.5688\n",
            "Epoch 97/100\n",
            "10/10 [==============================] - 1s 106ms/step - loss: 0.4593 - accuracy: 0.8010 - val_loss: 1.1025 - val_accuracy: 0.5531\n",
            "Epoch 98/100\n",
            "10/10 [==============================] - 1s 119ms/step - loss: 0.4544 - accuracy: 0.8052 - val_loss: 1.1585 - val_accuracy: 0.6031\n",
            "Epoch 99/100\n",
            "10/10 [==============================] - 1s 130ms/step - loss: 0.5891 - accuracy: 0.7781 - val_loss: 1.3255 - val_accuracy: 0.4875\n",
            "Epoch 100/100\n",
            "10/10 [==============================] - 1s 130ms/step - loss: 0.5754 - accuracy: 0.7635 - val_loss: 1.6276 - val_accuracy: 0.4625\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fa700020cd0>"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_true = np.array(y_test)\n",
        "y_pred = model.predict(data_testing)\n",
        "\n",
        "y_pred = np.squeeze(y_pred)\n",
        "y_pred = np.array(y_pred >= 0.5, dtype=np.int)\n",
        "\n",
        "from sklearn.metrics import classification_report\n",
        "print(classification_report(y_test, y_pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0dmnRtuQtLRp",
        "outputId": "72356ea7-8587-48fd-f0fd-734a400bedc6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.27      0.60      0.37        89\n",
            "           1       0.70      0.44      0.54       231\n",
            "\n",
            "   micro avg       0.45      0.48      0.47       320\n",
            "   macro avg       0.49      0.52      0.46       320\n",
            "weighted avg       0.58      0.48      0.49       320\n",
            " samples avg       0.43      0.48      0.45       320\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Valance"
      ],
      "metadata": {
        "id": "DVxrhuOlsJ_L"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# LSTM Examples: https://keras.io/examples/nlp/bidirectional_lstm_imdb/\n",
        "# https://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/ \n",
        "# LSTM for international airline passengers problem with time step regression framing\n",
        "import numpy\n",
        "import matplotlib.pyplot as plt\n",
        "from pandas import read_csv\n",
        "import math\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers import LSTM\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.metrics import mean_squared_error\n",
        "\n",
        "# reshape input to be [samples, time steps, features]\n",
        "#trainX = numpy.reshape(trainX, (trainX.shape[0], trainX.shape[1], 1))\n",
        "#testX = numpy.reshape(testX, (testX.shape[0], testX.shape[1], 1))\n",
        "\n",
        "# create and fit the LSTM network\n",
        "model = Sequential()\n",
        "model.add(LSTM(100, input_shape=(40, 8064), return_sequences=True))\n",
        "model.add(LSTM(100))\n",
        "model.add(Dense(20))\n",
        "model.add(Dropout(0.2))\n",
        "model.add(Dense(2))\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=[\"accuracy\"])\n",
        "model.fit(data_training, y_train, epochs=100, batch_size=100, verbose=1,validation_data=(data_testing, y_test))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nCpJuFv3reJ7",
        "outputId": "9ebb41d5-8e84-4365-d09d-cc7cf7509855"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "10/10 [==============================] - 4s 190ms/step - loss: 2.5329 - accuracy: 0.5990 - val_loss: 6.1735 - val_accuracy: 0.2531\n",
            "Epoch 2/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 1.3864 - accuracy: 0.5990 - val_loss: 1.8227 - val_accuracy: 0.3375\n",
            "Epoch 3/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.9035 - accuracy: 0.6344 - val_loss: 0.7767 - val_accuracy: 0.3594\n",
            "Epoch 4/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.7745 - accuracy: 0.5906 - val_loss: 0.8804 - val_accuracy: 0.2875\n",
            "Epoch 5/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.6565 - accuracy: 0.6604 - val_loss: 0.9272 - val_accuracy: 0.2625\n",
            "Epoch 6/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.6369 - accuracy: 0.6729 - val_loss: 1.0287 - val_accuracy: 0.2531\n",
            "Epoch 7/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.5993 - accuracy: 0.7010 - val_loss: 0.9221 - val_accuracy: 0.2531\n",
            "Epoch 8/100\n",
            "10/10 [==============================] - 1s 84ms/step - loss: 0.5777 - accuracy: 0.7208 - val_loss: 0.9990 - val_accuracy: 0.2594\n",
            "Epoch 9/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.5865 - accuracy: 0.7177 - val_loss: 0.8388 - val_accuracy: 0.3125\n",
            "Epoch 10/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.5597 - accuracy: 0.7312 - val_loss: 1.0884 - val_accuracy: 0.2656\n",
            "Epoch 11/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.5287 - accuracy: 0.7396 - val_loss: 1.0475 - val_accuracy: 0.2719\n",
            "Epoch 12/100\n",
            "10/10 [==============================] - 1s 85ms/step - loss: 0.5076 - accuracy: 0.7542 - val_loss: 0.9691 - val_accuracy: 0.2719\n",
            "Epoch 13/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.4860 - accuracy: 0.7760 - val_loss: 1.2796 - val_accuracy: 0.2688\n",
            "Epoch 14/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.4523 - accuracy: 0.8031 - val_loss: 0.7763 - val_accuracy: 0.5125\n",
            "Epoch 15/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.5237 - accuracy: 0.7760 - val_loss: 0.9054 - val_accuracy: 0.4094\n",
            "Epoch 16/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.4492 - accuracy: 0.8083 - val_loss: 1.2568 - val_accuracy: 0.3500\n",
            "Epoch 17/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.4121 - accuracy: 0.8344 - val_loss: 1.3763 - val_accuracy: 0.3594\n",
            "Epoch 18/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.3682 - accuracy: 0.8552 - val_loss: 2.1442 - val_accuracy: 0.3406\n",
            "Epoch 19/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.3162 - accuracy: 0.8823 - val_loss: 1.5257 - val_accuracy: 0.4812\n",
            "Epoch 20/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.3625 - accuracy: 0.8698 - val_loss: 4.0292 - val_accuracy: 0.2969\n",
            "Epoch 21/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.2613 - accuracy: 0.8969 - val_loss: 3.0494 - val_accuracy: 0.3531\n",
            "Epoch 22/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.2389 - accuracy: 0.9240 - val_loss: 2.8094 - val_accuracy: 0.4313\n",
            "Epoch 23/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.1816 - accuracy: 0.9354 - val_loss: 3.5756 - val_accuracy: 0.4125\n",
            "Epoch 24/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.2223 - accuracy: 0.9312 - val_loss: 4.0472 - val_accuracy: 0.4219\n",
            "Epoch 25/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.1979 - accuracy: 0.9427 - val_loss: 4.1611 - val_accuracy: 0.3875\n",
            "Epoch 26/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.1475 - accuracy: 0.9542 - val_loss: 5.2462 - val_accuracy: 0.3969\n",
            "Epoch 27/100\n",
            "10/10 [==============================] - 1s 78ms/step - loss: 0.2467 - accuracy: 0.9240 - val_loss: 3.4943 - val_accuracy: 0.4656\n",
            "Epoch 28/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.2424 - accuracy: 0.9187 - val_loss: 4.3937 - val_accuracy: 0.4281\n",
            "Epoch 29/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.1739 - accuracy: 0.9344 - val_loss: 4.7807 - val_accuracy: 0.3844\n",
            "Epoch 30/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.2443 - accuracy: 0.9344 - val_loss: 6.5709 - val_accuracy: 0.3000\n",
            "Epoch 31/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.1878 - accuracy: 0.9563 - val_loss: 3.1605 - val_accuracy: 0.5000\n",
            "Epoch 32/100\n",
            "10/10 [==============================] - 1s 78ms/step - loss: 0.2932 - accuracy: 0.9198 - val_loss: 4.0701 - val_accuracy: 0.4469\n",
            "Epoch 33/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.3958 - accuracy: 0.8969 - val_loss: 6.7771 - val_accuracy: 0.3406\n",
            "Epoch 34/100\n",
            "10/10 [==============================] - 1s 84ms/step - loss: 0.7718 - accuracy: 0.8177 - val_loss: 2.4855 - val_accuracy: 0.6938\n",
            "Epoch 35/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 3.2696 - accuracy: 0.6781 - val_loss: 8.3548 - val_accuracy: 0.2531\n",
            "Epoch 36/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 2.5171 - accuracy: 0.6354 - val_loss: 7.8383 - val_accuracy: 0.2656\n",
            "Epoch 37/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 2.6777 - accuracy: 0.7000 - val_loss: 5.3016 - val_accuracy: 0.2937\n",
            "Epoch 38/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 1.3611 - accuracy: 0.6646 - val_loss: 0.9581 - val_accuracy: 0.4906\n",
            "Epoch 39/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.8836 - accuracy: 0.7063 - val_loss: 1.3498 - val_accuracy: 0.3500\n",
            "Epoch 40/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.5594 - accuracy: 0.7531 - val_loss: 1.0668 - val_accuracy: 0.4187\n",
            "Epoch 41/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.5064 - accuracy: 0.7781 - val_loss: 1.3781 - val_accuracy: 0.3438\n",
            "Epoch 42/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.4647 - accuracy: 0.8042 - val_loss: 1.3185 - val_accuracy: 0.3656\n",
            "Epoch 43/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.4272 - accuracy: 0.8083 - val_loss: 1.4024 - val_accuracy: 0.3469\n",
            "Epoch 44/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.3948 - accuracy: 0.8427 - val_loss: 1.3634 - val_accuracy: 0.3562\n",
            "Epoch 45/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.3547 - accuracy: 0.8469 - val_loss: 1.7354 - val_accuracy: 0.3562\n",
            "Epoch 46/100\n",
            "10/10 [==============================] - 1s 86ms/step - loss: 0.3376 - accuracy: 0.8698 - val_loss: 1.9547 - val_accuracy: 0.3719\n",
            "Epoch 47/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.3082 - accuracy: 0.8865 - val_loss: 2.3001 - val_accuracy: 0.3562\n",
            "Epoch 48/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.2783 - accuracy: 0.8917 - val_loss: 1.6865 - val_accuracy: 0.4281\n",
            "Epoch 49/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.2409 - accuracy: 0.9198 - val_loss: 3.2955 - val_accuracy: 0.3406\n",
            "Epoch 50/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.2237 - accuracy: 0.9167 - val_loss: 2.2377 - val_accuracy: 0.4812\n",
            "Epoch 51/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.1996 - accuracy: 0.9250 - val_loss: 3.1264 - val_accuracy: 0.4281\n",
            "Epoch 52/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.1621 - accuracy: 0.9427 - val_loss: 3.8379 - val_accuracy: 0.3719\n",
            "Epoch 53/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.1500 - accuracy: 0.9396 - val_loss: 3.9690 - val_accuracy: 0.3750\n",
            "Epoch 54/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.1462 - accuracy: 0.9438 - val_loss: 3.9158 - val_accuracy: 0.4187\n",
            "Epoch 55/100\n",
            "10/10 [==============================] - 1s 84ms/step - loss: 0.1125 - accuracy: 0.9667 - val_loss: 4.0033 - val_accuracy: 0.4375\n",
            "Epoch 56/100\n",
            "10/10 [==============================] - 1s 84ms/step - loss: 0.0991 - accuracy: 0.9698 - val_loss: 4.2277 - val_accuracy: 0.4125\n",
            "Epoch 57/100\n",
            "10/10 [==============================] - 1s 78ms/step - loss: 0.0836 - accuracy: 0.9781 - val_loss: 5.2864 - val_accuracy: 0.3781\n",
            "Epoch 58/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.5242 - accuracy: 0.9062 - val_loss: 6.8524 - val_accuracy: 0.3094\n",
            "Epoch 59/100\n",
            "10/10 [==============================] - 1s 86ms/step - loss: 0.5520 - accuracy: 0.8750 - val_loss: 4.4413 - val_accuracy: 0.3906\n",
            "Epoch 60/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.8656 - accuracy: 0.8104 - val_loss: 3.9399 - val_accuracy: 0.4094\n",
            "Epoch 61/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.6310 - accuracy: 0.8479 - val_loss: 4.0175 - val_accuracy: 0.3812\n",
            "Epoch 62/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.8425 - accuracy: 0.8365 - val_loss: 3.9948 - val_accuracy: 0.4094\n",
            "Epoch 63/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.6515 - accuracy: 0.8031 - val_loss: 4.8205 - val_accuracy: 0.3125\n",
            "Epoch 64/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.9390 - accuracy: 0.7531 - val_loss: 2.0895 - val_accuracy: 0.4469\n",
            "Epoch 65/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.6395 - accuracy: 0.8021 - val_loss: 2.6000 - val_accuracy: 0.3531\n",
            "Epoch 66/100\n",
            "10/10 [==============================] - 1s 85ms/step - loss: 0.4232 - accuracy: 0.8406 - val_loss: 1.5831 - val_accuracy: 0.4469\n",
            "Epoch 67/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.3754 - accuracy: 0.8542 - val_loss: 2.7395 - val_accuracy: 0.3344\n",
            "Epoch 68/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.2950 - accuracy: 0.8958 - val_loss: 2.0893 - val_accuracy: 0.4437\n",
            "Epoch 69/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.2725 - accuracy: 0.9021 - val_loss: 2.6888 - val_accuracy: 0.3844\n",
            "Epoch 70/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.2365 - accuracy: 0.9354 - val_loss: 2.8439 - val_accuracy: 0.3906\n",
            "Epoch 71/100\n",
            "10/10 [==============================] - 1s 84ms/step - loss: 0.1894 - accuracy: 0.9292 - val_loss: 3.0353 - val_accuracy: 0.4000\n",
            "Epoch 72/100\n",
            "10/10 [==============================] - 1s 84ms/step - loss: 0.1639 - accuracy: 0.9490 - val_loss: 3.4033 - val_accuracy: 0.3844\n",
            "Epoch 73/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.1383 - accuracy: 0.9510 - val_loss: 3.2820 - val_accuracy: 0.4187\n",
            "Epoch 74/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.1194 - accuracy: 0.9594 - val_loss: 3.5675 - val_accuracy: 0.3906\n",
            "Epoch 75/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.0989 - accuracy: 0.9688 - val_loss: 4.1530 - val_accuracy: 0.4125\n",
            "Epoch 76/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.0865 - accuracy: 0.9740 - val_loss: 3.9086 - val_accuracy: 0.4313\n",
            "Epoch 77/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.0798 - accuracy: 0.9740 - val_loss: 4.5138 - val_accuracy: 0.4156\n",
            "Epoch 78/100\n",
            "10/10 [==============================] - 1s 100ms/step - loss: 0.0674 - accuracy: 0.9802 - val_loss: 3.9272 - val_accuracy: 0.4656\n",
            "Epoch 79/100\n",
            "10/10 [==============================] - 1s 108ms/step - loss: 0.0623 - accuracy: 0.9781 - val_loss: 4.5735 - val_accuracy: 0.4375\n",
            "Epoch 80/100\n",
            "10/10 [==============================] - 1s 103ms/step - loss: 0.0522 - accuracy: 0.9865 - val_loss: 5.2638 - val_accuracy: 0.4094\n",
            "Epoch 81/100\n",
            "10/10 [==============================] - 1s 87ms/step - loss: 0.0443 - accuracy: 0.9875 - val_loss: 4.5620 - val_accuracy: 0.4719\n",
            "Epoch 82/100\n",
            "10/10 [==============================] - 1s 79ms/step - loss: 0.0404 - accuracy: 0.9906 - val_loss: 4.7004 - val_accuracy: 0.4656\n",
            "Epoch 83/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.0373 - accuracy: 0.9906 - val_loss: 5.3842 - val_accuracy: 0.4000\n",
            "Epoch 84/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.0320 - accuracy: 0.9917 - val_loss: 5.2405 - val_accuracy: 0.4219\n",
            "Epoch 85/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.0473 - accuracy: 0.9885 - val_loss: 5.6529 - val_accuracy: 0.4250\n",
            "Epoch 86/100\n",
            "10/10 [==============================] - 1s 82ms/step - loss: 0.0495 - accuracy: 0.9896 - val_loss: 5.1602 - val_accuracy: 0.4281\n",
            "Epoch 87/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.0330 - accuracy: 0.9896 - val_loss: 5.1078 - val_accuracy: 0.4531\n",
            "Epoch 88/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.0287 - accuracy: 0.9927 - val_loss: 6.1487 - val_accuracy: 0.4062\n",
            "Epoch 89/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.0265 - accuracy: 0.9906 - val_loss: 5.0532 - val_accuracy: 0.4500\n",
            "Epoch 90/100\n",
            "10/10 [==============================] - 1s 83ms/step - loss: 0.0251 - accuracy: 0.9917 - val_loss: 5.2140 - val_accuracy: 0.4406\n",
            "Epoch 91/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.0216 - accuracy: 0.9948 - val_loss: 6.1625 - val_accuracy: 0.3781\n",
            "Epoch 92/100\n",
            "10/10 [==============================] - 1s 78ms/step - loss: 0.0227 - accuracy: 0.9927 - val_loss: 6.7256 - val_accuracy: 0.3688\n",
            "Epoch 93/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.0189 - accuracy: 0.9948 - val_loss: 5.7098 - val_accuracy: 0.4437\n",
            "Epoch 94/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.0156 - accuracy: 0.9958 - val_loss: 6.6717 - val_accuracy: 0.3875\n",
            "Epoch 95/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.0163 - accuracy: 0.9948 - val_loss: 6.3631 - val_accuracy: 0.3969\n",
            "Epoch 96/100\n",
            "10/10 [==============================] - 1s 80ms/step - loss: 0.0125 - accuracy: 0.9990 - val_loss: 6.2418 - val_accuracy: 0.4125\n",
            "Epoch 97/100\n",
            "10/10 [==============================] - 1s 81ms/step - loss: 0.0122 - accuracy: 0.9979 - val_loss: 6.1264 - val_accuracy: 0.4250\n",
            "Epoch 98/100\n",
            "10/10 [==============================] - 1s 104ms/step - loss: 0.0116 - accuracy: 0.9979 - val_loss: 5.9467 - val_accuracy: 0.4406\n",
            "Epoch 99/100\n",
            "10/10 [==============================] - 1s 95ms/step - loss: 0.0080 - accuracy: 1.0000 - val_loss: 6.6474 - val_accuracy: 0.4062\n",
            "Epoch 100/100\n",
            "10/10 [==============================] - 1s 91ms/step - loss: 0.0075 - accuracy: 0.9990 - val_loss: 5.2985 - val_accuracy: 0.4719\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fe6979a5dd0>"
            ]
          },
          "metadata": {},
          "execution_count": 75
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Bi-LSTM"
      ],
      "metadata": {
        "id": "fv5YGMRnREBP"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Arousal"
      ],
      "metadata": {
        "id": "IHDIWr9ftgPh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# LSTM Examples: https://keras.io/examples/nlp/bidirectional_lstm_imdb/\n",
        "# https://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/ \n",
        "# LSTM for international airline passengers problem with time step regression framing\n",
        "# Bi-LSTM https://analyticsindiamag.com/complete-guide-to-bidirectional-lstm-with-python-codes/#:~:text=Bidirectional%20long%2Dshort%20term%20memory(bi%2Dlstm)%20is,different%20from%20the%20regular%20LSTM.\n",
        "import numpy\n",
        "import matplotlib.pyplot as plt\n",
        "from pandas import read_csv\n",
        "import math\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers import LSTM\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from keras.layers import Dropout, Embedding, Bidirectional\n",
        "\n",
        "# reshape input to be [samples, time steps, features]\n",
        "#trainX = numpy.reshape(trainX, (trainX.shape[0], trainX.shape[1], 1))\n",
        "#testX = numpy.reshape(testX, (testX.shape[0], testX.shape[1], 1))\n",
        "\n",
        "# create and fit the LSTM network\n",
        "model = Sequential()\n",
        "model.add(Bidirectional(LSTM(100, input_shape=(40,8064), return_sequences=True)))\n",
        "model.add(Bidirectional(LSTM(100)))\n",
        "model.add(Dense(20))\n",
        "model.add(Dropout(0.2))\n",
        "model.add(Dense(2))\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=[\"accuracy\"])\n",
        "model.fit(data_training, y_train, epochs=100, batch_size=100, verbose=1,validation_data=(data_testing, y_test))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "upCi92OhthkY",
        "outputId": "14482ad0-75e1-435d-95a0-5e4d9f2f3f26"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "10/10 [==============================] - 10s 361ms/step - loss: 7.6063 - accuracy: 0.4542 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 2/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.3555 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 3/100\n",
            "10/10 [==============================] - 1s 134ms/step - loss: 8.3555 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 4/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.3555 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 5/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.3555 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 6/100\n",
            "10/10 [==============================] - 1s 140ms/step - loss: 8.3555 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 7/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 8.3555 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 8/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 8.3555 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 9/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 8.3434 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 10/100\n",
            "10/10 [==============================] - 1s 140ms/step - loss: 7.7689 - accuracy: 0.4552 - val_loss: 7.7125 - val_accuracy: 0.2781\n",
            "Epoch 11/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 7.7047 - accuracy: 0.4833 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 12/100\n",
            "10/10 [==============================] - 1s 134ms/step - loss: 7.7125 - accuracy: 0.5500 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 13/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 7.7125 - accuracy: 0.5479 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 14/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 7.6989 - accuracy: 0.5458 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 15/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 7.2381 - accuracy: 0.5448 - val_loss: 2.5226 - val_accuracy: 0.7219\n",
            "Epoch 16/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.7368 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 17/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 6.7272 - accuracy: 0.5448 - val_loss: 7.6136 - val_accuracy: 0.7219\n",
            "Epoch 18/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.6537 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 19/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.6893 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 20/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.7859 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 21/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 6.8510 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 22/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.8675 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 23/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 6.6586 - accuracy: 0.5448 - val_loss: 4.2193 - val_accuracy: 0.7219\n",
            "Epoch 24/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 6.6532 - accuracy: 0.5448 - val_loss: 3.7382 - val_accuracy: 0.7219\n",
            "Epoch 25/100\n",
            "10/10 [==============================] - 1s 140ms/step - loss: 6.7416 - accuracy: 0.5448 - val_loss: 2.7591 - val_accuracy: 0.7219\n",
            "Epoch 26/100\n",
            "10/10 [==============================] - 1s 140ms/step - loss: 6.6142 - accuracy: 0.5448 - val_loss: 2.5684 - val_accuracy: 0.7219\n",
            "Epoch 27/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 6.3906 - accuracy: 0.5448 - val_loss: 2.6749 - val_accuracy: 0.7219\n",
            "Epoch 28/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.5582 - accuracy: 0.5448 - val_loss: 2.6229 - val_accuracy: 0.7219\n",
            "Epoch 29/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 6.4072 - accuracy: 0.5448 - val_loss: 3.9178 - val_accuracy: 0.7219\n",
            "Epoch 30/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.5971 - accuracy: 0.5448 - val_loss: 3.3960 - val_accuracy: 0.7219\n",
            "Epoch 31/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.5009 - accuracy: 0.5448 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 32/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 7.0061 - accuracy: 0.5448 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 33/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.9962 - accuracy: 0.5448 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 34/100\n",
            "10/10 [==============================] - 1s 135ms/step - loss: 6.7700 - accuracy: 0.5448 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 35/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.7581 - accuracy: 0.5458 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 36/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 6.9225 - accuracy: 0.5448 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 37/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 7.3642 - accuracy: 0.5437 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 38/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 7.6454 - accuracy: 0.5458 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 39/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 7.7143 - accuracy: 0.5406 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 40/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 7.6979 - accuracy: 0.5365 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 41/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 7.7010 - accuracy: 0.5479 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 42/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 7.7050 - accuracy: 0.5375 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 43/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 7.6976 - accuracy: 0.5406 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 44/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 7.7134 - accuracy: 0.5469 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 45/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 7.6970 - accuracy: 0.5479 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 46/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 7.6978 - accuracy: 0.5469 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 47/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 7.7127 - accuracy: 0.5406 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 48/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 7.6990 - accuracy: 0.5344 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 49/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 7.6940 - accuracy: 0.5365 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 50/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 7.6663 - accuracy: 0.5437 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 51/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 7.5777 - accuracy: 0.5427 - val_loss: 7.7125 - val_accuracy: 0.7219\n",
            "Epoch 52/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 7.1843 - accuracy: 0.5448 - val_loss: 7.1683 - val_accuracy: 0.7219\n",
            "Epoch 53/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 6.6383 - accuracy: 0.5448 - val_loss: 3.3361 - val_accuracy: 0.7219\n",
            "Epoch 54/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.7765 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 55/100\n",
            "10/10 [==============================] - 1s 140ms/step - loss: 6.5670 - accuracy: 0.5448 - val_loss: 2.4710 - val_accuracy: 0.7219\n",
            "Epoch 56/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 6.4091 - accuracy: 0.5458 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 57/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.9692 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 58/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.9817 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 59/100\n",
            "10/10 [==============================] - 2s 182ms/step - loss: 6.9816 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 60/100\n",
            "10/10 [==============================] - 1s 140ms/step - loss: 6.9896 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 61/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.9750 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 62/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.9676 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 63/100\n",
            "10/10 [==============================] - 1s 135ms/step - loss: 6.9743 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 64/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 6.9690 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 65/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.9255 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 66/100\n",
            "10/10 [==============================] - 1s 135ms/step - loss: 6.8883 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 67/100\n",
            "10/10 [==============================] - 1s 135ms/step - loss: 6.9052 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 68/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 6.8686 - accuracy: 0.5437 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 69/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.8835 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 70/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.9308 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 71/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 6.9359 - accuracy: 0.5427 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 72/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.8950 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 73/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.9062 - accuracy: 0.5448 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 74/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 6.8396 - accuracy: 0.5427 - val_loss: 4.2656 - val_accuracy: 0.7219\n",
            "Epoch 75/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 6.5805 - accuracy: 0.5406 - val_loss: 7.1526 - val_accuracy: 0.7219\n",
            "Epoch 76/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 7.3614 - accuracy: 0.5385 - val_loss: 7.7125 - val_accuracy: 0.2781\n",
            "Epoch 77/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 7.5927 - accuracy: 0.4365 - val_loss: 7.7125 - val_accuracy: 0.2781\n",
            "Epoch 78/100\n",
            "10/10 [==============================] - 1s 135ms/step - loss: 7.5731 - accuracy: 0.4656 - val_loss: 7.5648 - val_accuracy: 0.2781\n",
            "Epoch 79/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 7.6742 - accuracy: 0.4521 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 80/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.1947 - accuracy: 0.4563 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 81/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 8.1858 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 82/100\n",
            "10/10 [==============================] - 2s 156ms/step - loss: 8.2882 - accuracy: 0.4563 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 83/100\n",
            "10/10 [==============================] - 2s 155ms/step - loss: 7.9867 - accuracy: 0.4573 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 84/100\n",
            "10/10 [==============================] - 1s 150ms/step - loss: 8.0118 - accuracy: 0.4615 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 85/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.1645 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 86/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 8.0413 - accuracy: 0.4542 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 87/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 8.0887 - accuracy: 0.4563 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 88/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 8.0496 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 89/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 7.9327 - accuracy: 0.4573 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 90/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 8.0370 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 91/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.1422 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 92/100\n",
            "10/10 [==============================] - 1s 140ms/step - loss: 8.0621 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 93/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 8.1734 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 94/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 8.0040 - accuracy: 0.4542 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 95/100\n",
            "10/10 [==============================] - 1s 139ms/step - loss: 8.0559 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 96/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.1155 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 97/100\n",
            "10/10 [==============================] - 1s 138ms/step - loss: 8.1245 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 98/100\n",
            "10/10 [==============================] - 1s 141ms/step - loss: 8.0918 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 99/100\n",
            "10/10 [==============================] - 1s 137ms/step - loss: 8.0748 - accuracy: 0.4542 - val_loss: 11.0715 - val_accuracy: 0.2781\n",
            "Epoch 100/100\n",
            "10/10 [==============================] - 1s 136ms/step - loss: 8.0510 - accuracy: 0.4552 - val_loss: 11.0715 - val_accuracy: 0.2781\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fa69358ff10>"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Valance"
      ],
      "metadata": {
        "id": "7iL8Wat0teW3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# LSTM Examples: https://keras.io/examples/nlp/bidirectional_lstm_imdb/\n",
        "# https://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/ \n",
        "# LSTM for international airline passengers problem with time step regression framing\n",
        "# Bi-LSTM https://analyticsindiamag.com/complete-guide-to-bidirectional-lstm-with-python-codes/#:~:text=Bidirectional%20long%2Dshort%20term%20memory(bi%2Dlstm)%20is,different%20from%20the%20regular%20LSTM.\n",
        "import numpy\n",
        "import matplotlib.pyplot as plt\n",
        "from pandas import read_csv\n",
        "import math\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers import LSTM\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from keras.layers import Dropout, Embedding, Bidirectional\n",
        "\n",
        "# reshape input to be [samples, time steps, features]\n",
        "#trainX = numpy.reshape(trainX, (trainX.shape[0], trainX.shape[1], 1))\n",
        "#testX = numpy.reshape(testX, (testX.shape[0], testX.shape[1], 1))\n",
        "\n",
        "# create and fit the LSTM network\n",
        "model = Sequential()\n",
        "model.add(Bidirectional(LSTM(100, input_shape=(40,8064), return_sequences=True)))\n",
        "model.add(Bidirectional(LSTM(100)))\n",
        "model.add(Dense(20))\n",
        "model.add(Dropout(0.2))\n",
        "model.add(Dense(2))\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=[\"accuracy\"])\n",
        "model.fit(data_training, y_train, epochs=100, batch_size=100, verbose=1,validation_data=(data_testing, y_test))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iWlwsfqztpZ9",
        "outputId": "cc21a65b-8deb-4038-ea46-b8988159a6b7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "10/10 [==============================] - 9s 285ms/step - loss: 3.9141 - accuracy: 0.5917 - val_loss: 6.1383 - val_accuracy: 0.3594\n",
            "Epoch 2/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 3.4115 - accuracy: 0.5771 - val_loss: 5.5747 - val_accuracy: 0.3750\n",
            "Epoch 3/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 3.1473 - accuracy: 0.6552 - val_loss: 10.2296 - val_accuracy: 0.2562\n",
            "Epoch 4/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 4.2789 - accuracy: 0.6646 - val_loss: 9.5081 - val_accuracy: 0.2656\n",
            "Epoch 5/100\n",
            "10/10 [==============================] - 1s 110ms/step - loss: 4.0768 - accuracy: 0.6583 - val_loss: 8.0533 - val_accuracy: 0.2812\n",
            "Epoch 6/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 3.7445 - accuracy: 0.6562 - val_loss: 8.0677 - val_accuracy: 0.2719\n",
            "Epoch 7/100\n",
            "10/10 [==============================] - 1s 110ms/step - loss: 3.7225 - accuracy: 0.6740 - val_loss: 8.6064 - val_accuracy: 0.2719\n",
            "Epoch 8/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 3.5326 - accuracy: 0.6500 - val_loss: 5.7896 - val_accuracy: 0.4125\n",
            "Epoch 9/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 3.7370 - accuracy: 0.5771 - val_loss: 3.2434 - val_accuracy: 0.4875\n",
            "Epoch 10/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 2.9700 - accuracy: 0.5292 - val_loss: 4.5194 - val_accuracy: 0.4000\n",
            "Epoch 11/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 2.4577 - accuracy: 0.6344 - val_loss: 4.6305 - val_accuracy: 0.5219\n",
            "Epoch 12/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 3.0978 - accuracy: 0.5948 - val_loss: 4.5131 - val_accuracy: 0.3938\n",
            "Epoch 13/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 3.2588 - accuracy: 0.5615 - val_loss: 1.9520 - val_accuracy: 0.5844\n",
            "Epoch 14/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 2.0143 - accuracy: 0.6135 - val_loss: 4.3713 - val_accuracy: 0.2906\n",
            "Epoch 15/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 2.2266 - accuracy: 0.6260 - val_loss: 2.0354 - val_accuracy: 0.6375\n",
            "Epoch 16/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 3.8869 - accuracy: 0.4823 - val_loss: 2.7635 - val_accuracy: 0.5437\n",
            "Epoch 17/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 4.0034 - accuracy: 0.4802 - val_loss: 2.1290 - val_accuracy: 0.6438\n",
            "Epoch 18/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 3.9201 - accuracy: 0.4885 - val_loss: 2.4419 - val_accuracy: 0.5625\n",
            "Epoch 19/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 3.9140 - accuracy: 0.5177 - val_loss: 2.3571 - val_accuracy: 0.5500\n",
            "Epoch 20/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 3.6745 - accuracy: 0.5323 - val_loss: 2.4705 - val_accuracy: 0.5375\n",
            "Epoch 21/100\n",
            "10/10 [==============================] - 1s 118ms/step - loss: 3.7571 - accuracy: 0.5344 - val_loss: 2.4735 - val_accuracy: 0.5344\n",
            "Epoch 22/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 3.5456 - accuracy: 0.5427 - val_loss: 2.2321 - val_accuracy: 0.5562\n",
            "Epoch 23/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 3.3530 - accuracy: 0.5562 - val_loss: 2.2590 - val_accuracy: 0.5500\n",
            "Epoch 24/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 3.4127 - accuracy: 0.5656 - val_loss: 2.2078 - val_accuracy: 0.5437\n",
            "Epoch 25/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 3.1385 - accuracy: 0.5708 - val_loss: 2.6546 - val_accuracy: 0.5219\n",
            "Epoch 26/100\n",
            "10/10 [==============================] - 1s 117ms/step - loss: 3.0310 - accuracy: 0.5896 - val_loss: 2.0895 - val_accuracy: 0.5312\n",
            "Epoch 27/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 2.8531 - accuracy: 0.5875 - val_loss: 2.2998 - val_accuracy: 0.4844\n",
            "Epoch 28/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 1.7312 - accuracy: 0.6427 - val_loss: 2.2818 - val_accuracy: 0.3000\n",
            "Epoch 29/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.8274 - accuracy: 0.7167 - val_loss: 1.1498 - val_accuracy: 0.4469\n",
            "Epoch 30/100\n",
            "10/10 [==============================] - 1s 117ms/step - loss: 0.5900 - accuracy: 0.7188 - val_loss: 1.7515 - val_accuracy: 0.2969\n",
            "Epoch 31/100\n",
            "10/10 [==============================] - 1s 124ms/step - loss: 0.5527 - accuracy: 0.7500 - val_loss: 1.1254 - val_accuracy: 0.3313\n",
            "Epoch 32/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.4886 - accuracy: 0.8062 - val_loss: 1.7111 - val_accuracy: 0.3125\n",
            "Epoch 33/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.4804 - accuracy: 0.7885 - val_loss: 1.0516 - val_accuracy: 0.4125\n",
            "Epoch 34/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.4684 - accuracy: 0.7854 - val_loss: 2.1982 - val_accuracy: 0.3063\n",
            "Epoch 35/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.4288 - accuracy: 0.8271 - val_loss: 1.4227 - val_accuracy: 0.3438\n",
            "Epoch 36/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 0.4013 - accuracy: 0.8302 - val_loss: 1.9119 - val_accuracy: 0.3125\n",
            "Epoch 37/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.3727 - accuracy: 0.8354 - val_loss: 1.6678 - val_accuracy: 0.3531\n",
            "Epoch 38/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.3527 - accuracy: 0.8646 - val_loss: 2.2703 - val_accuracy: 0.3187\n",
            "Epoch 39/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.3405 - accuracy: 0.8562 - val_loss: 2.0885 - val_accuracy: 0.3313\n",
            "Epoch 40/100\n",
            "10/10 [==============================] - 1s 118ms/step - loss: 0.3281 - accuracy: 0.8969 - val_loss: 3.0608 - val_accuracy: 0.3000\n",
            "Epoch 41/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.3042 - accuracy: 0.8583 - val_loss: 1.8670 - val_accuracy: 0.3938\n",
            "Epoch 42/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.2876 - accuracy: 0.9094 - val_loss: 3.2761 - val_accuracy: 0.3125\n",
            "Epoch 43/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 0.2616 - accuracy: 0.8938 - val_loss: 2.6835 - val_accuracy: 0.3656\n",
            "Epoch 44/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 0.2428 - accuracy: 0.9438 - val_loss: 3.4328 - val_accuracy: 0.3156\n",
            "Epoch 45/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.2302 - accuracy: 0.9198 - val_loss: 2.7674 - val_accuracy: 0.3938\n",
            "Epoch 46/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.2094 - accuracy: 0.9260 - val_loss: 3.2528 - val_accuracy: 0.3938\n",
            "Epoch 47/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 0.6581 - accuracy: 0.8771 - val_loss: 8.5446 - val_accuracy: 0.2594\n",
            "Epoch 48/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 2.3854 - accuracy: 0.7271 - val_loss: 5.3172 - val_accuracy: 0.3438\n",
            "Epoch 49/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 2.8835 - accuracy: 0.7365 - val_loss: 6.3218 - val_accuracy: 0.3000\n",
            "Epoch 50/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 2.1614 - accuracy: 0.7396 - val_loss: 3.6322 - val_accuracy: 0.3906\n",
            "Epoch 51/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 1.8859 - accuracy: 0.7490 - val_loss: 5.1220 - val_accuracy: 0.2781\n",
            "Epoch 52/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 1.2942 - accuracy: 0.7479 - val_loss: 1.4113 - val_accuracy: 0.4719\n",
            "Epoch 53/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.5251 - accuracy: 0.7885 - val_loss: 1.8454 - val_accuracy: 0.3938\n",
            "Epoch 54/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.5003 - accuracy: 0.7865 - val_loss: 2.3628 - val_accuracy: 0.3500\n",
            "Epoch 55/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.4349 - accuracy: 0.8188 - val_loss: 2.5793 - val_accuracy: 0.4594\n",
            "Epoch 56/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.3816 - accuracy: 0.8365 - val_loss: 2.3611 - val_accuracy: 0.3406\n",
            "Epoch 57/100\n",
            "10/10 [==============================] - 1s 117ms/step - loss: 0.3718 - accuracy: 0.8604 - val_loss: 2.4527 - val_accuracy: 0.3562\n",
            "Epoch 58/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.3220 - accuracy: 0.8719 - val_loss: 2.4858 - val_accuracy: 0.3781\n",
            "Epoch 59/100\n",
            "10/10 [==============================] - 1s 119ms/step - loss: 0.3070 - accuracy: 0.8948 - val_loss: 2.9159 - val_accuracy: 0.3469\n",
            "Epoch 60/100\n",
            "10/10 [==============================] - 1s 154ms/step - loss: 0.2842 - accuracy: 0.8896 - val_loss: 2.6861 - val_accuracy: 0.3719\n",
            "Epoch 61/100\n",
            "10/10 [==============================] - 1s 144ms/step - loss: 0.2733 - accuracy: 0.9250 - val_loss: 3.4216 - val_accuracy: 0.3375\n",
            "Epoch 62/100\n",
            "10/10 [==============================] - 1s 142ms/step - loss: 0.2505 - accuracy: 0.8938 - val_loss: 2.9071 - val_accuracy: 0.3812\n",
            "Epoch 63/100\n",
            "10/10 [==============================] - 1s 123ms/step - loss: 0.2370 - accuracy: 0.9365 - val_loss: 3.7095 - val_accuracy: 0.3469\n",
            "Epoch 64/100\n",
            "10/10 [==============================] - 1s 120ms/step - loss: 0.2241 - accuracy: 0.9115 - val_loss: 3.2179 - val_accuracy: 0.3688\n",
            "Epoch 65/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 0.2072 - accuracy: 0.9521 - val_loss: 4.0867 - val_accuracy: 0.3313\n",
            "Epoch 66/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.1991 - accuracy: 0.9417 - val_loss: 3.5450 - val_accuracy: 0.3875\n",
            "Epoch 67/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 0.1856 - accuracy: 0.9500 - val_loss: 3.6679 - val_accuracy: 0.3812\n",
            "Epoch 68/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 0.1656 - accuracy: 0.9531 - val_loss: 4.0567 - val_accuracy: 0.3469\n",
            "Epoch 69/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 0.1592 - accuracy: 0.9594 - val_loss: 4.4544 - val_accuracy: 0.3531\n",
            "Epoch 70/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 0.1996 - accuracy: 0.9479 - val_loss: 6.6204 - val_accuracy: 0.3000\n",
            "Epoch 71/100\n",
            "10/10 [==============================] - 1s 124ms/step - loss: 0.7798 - accuracy: 0.8729 - val_loss: 5.0907 - val_accuracy: 0.3812\n",
            "Epoch 72/100\n",
            "10/10 [==============================] - 1s 121ms/step - loss: 0.6321 - accuracy: 0.9021 - val_loss: 4.0520 - val_accuracy: 0.4187\n",
            "Epoch 73/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 0.4691 - accuracy: 0.9177 - val_loss: 4.3419 - val_accuracy: 0.3469\n",
            "Epoch 74/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 0.3499 - accuracy: 0.9396 - val_loss: 4.2949 - val_accuracy: 0.3406\n",
            "Epoch 75/100\n",
            "10/10 [==============================] - 1s 110ms/step - loss: 0.2117 - accuracy: 0.9417 - val_loss: 4.0247 - val_accuracy: 0.4031\n",
            "Epoch 76/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 0.1268 - accuracy: 0.9719 - val_loss: 4.3189 - val_accuracy: 0.3625\n",
            "Epoch 77/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.1029 - accuracy: 0.9802 - val_loss: 3.7840 - val_accuracy: 0.4219\n",
            "Epoch 78/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.1056 - accuracy: 0.9708 - val_loss: 3.1851 - val_accuracy: 0.4812\n",
            "Epoch 79/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.0973 - accuracy: 0.9677 - val_loss: 4.6078 - val_accuracy: 0.4125\n",
            "Epoch 80/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.0982 - accuracy: 0.9750 - val_loss: 4.1978 - val_accuracy: 0.3875\n",
            "Epoch 81/100\n",
            "10/10 [==============================] - 1s 110ms/step - loss: 0.1026 - accuracy: 0.9708 - val_loss: 4.2179 - val_accuracy: 0.4531\n",
            "Epoch 82/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 0.3638 - accuracy: 0.9135 - val_loss: 2.0535 - val_accuracy: 0.6156\n",
            "Epoch 83/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 2.0895 - accuracy: 0.7781 - val_loss: 8.8387 - val_accuracy: 0.2812\n",
            "Epoch 84/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 1.2235 - accuracy: 0.8052 - val_loss: 4.5248 - val_accuracy: 0.4094\n",
            "Epoch 85/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.7804 - accuracy: 0.8510 - val_loss: 3.1905 - val_accuracy: 0.3781\n",
            "Epoch 86/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.6672 - accuracy: 0.8677 - val_loss: 2.5792 - val_accuracy: 0.5219\n",
            "Epoch 87/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 0.6640 - accuracy: 0.8490 - val_loss: 3.5244 - val_accuracy: 0.4000\n",
            "Epoch 88/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 1.4023 - accuracy: 0.7188 - val_loss: 5.8742 - val_accuracy: 0.2844\n",
            "Epoch 89/100\n",
            "10/10 [==============================] - 1s 131ms/step - loss: 1.7832 - accuracy: 0.6854 - val_loss: 2.0774 - val_accuracy: 0.6375\n",
            "Epoch 90/100\n",
            "10/10 [==============================] - 1s 123ms/step - loss: 2.2268 - accuracy: 0.6865 - val_loss: 3.6962 - val_accuracy: 0.4875\n",
            "Epoch 91/100\n",
            "10/10 [==============================] - 1s 132ms/step - loss: 1.9876 - accuracy: 0.6687 - val_loss: 2.3766 - val_accuracy: 0.6375\n",
            "Epoch 92/100\n",
            "10/10 [==============================] - 1s 113ms/step - loss: 1.6338 - accuracy: 0.7115 - val_loss: 2.1236 - val_accuracy: 0.4437\n",
            "Epoch 93/100\n",
            "10/10 [==============================] - 1s 111ms/step - loss: 1.9293 - accuracy: 0.7250 - val_loss: 2.4197 - val_accuracy: 0.4500\n",
            "Epoch 94/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.9634 - accuracy: 0.7729 - val_loss: 3.4871 - val_accuracy: 0.3063\n",
            "Epoch 95/100\n",
            "10/10 [==============================] - 1s 119ms/step - loss: 0.4194 - accuracy: 0.8396 - val_loss: 1.9959 - val_accuracy: 0.4469\n",
            "Epoch 96/100\n",
            "10/10 [==============================] - 1s 112ms/step - loss: 0.3865 - accuracy: 0.8615 - val_loss: 2.5281 - val_accuracy: 0.3344\n",
            "Epoch 97/100\n",
            "10/10 [==============================] - 1s 114ms/step - loss: 0.3339 - accuracy: 0.8562 - val_loss: 3.0006 - val_accuracy: 0.3031\n",
            "Epoch 98/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.2918 - accuracy: 0.8760 - val_loss: 2.6188 - val_accuracy: 0.3438\n",
            "Epoch 99/100\n",
            "10/10 [==============================] - 1s 115ms/step - loss: 0.2600 - accuracy: 0.9094 - val_loss: 3.1740 - val_accuracy: 0.3281\n",
            "Epoch 100/100\n",
            "10/10 [==============================] - 1s 116ms/step - loss: 0.2462 - accuracy: 0.9052 - val_loss: 3.2674 - val_accuracy: 0.3281\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fe697e8ea90>"
            ]
          },
          "metadata": {},
          "execution_count": 80
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.rcParams[\"figure.figsize\"] = (20,8)\n",
        "plt.rcParams[\"font.size\"] = 20\n",
        "\n",
        "plt.plot(H.history['accuracy'])\n",
        "plt.plot(H.history['val_accuracy'])\n",
        "plt.title('LSTM')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train acc', 'test acc'], loc='upper left')\n",
        "plt.savefig('/content/drive/Shareddrives/DEAP EEG/MDPI Sensor Journal/Paper figure/Bi-LSTM 100 epoch train and test accuracy curve.png', dpi=300)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 393
        },
        "id": "vf1amAhXuNib",
        "outputId": "d7bdff7e-f8ba-4a59-cfef-2ad0ef9da780"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_true = np.array(y_test)\n",
        "y_pred = model.predict(data_testing)\n",
        "\n",
        "y_pred = np.squeeze(y_pred)\n",
        "y_pred = np.array(y_pred >= 0.5, dtype=np.int)\n",
        "\n",
        "from sklearn.metrics import classification_report\n",
        "print(classification_report(y_test, y_pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Sq-E3Q3UuWjb",
        "outputId": "77913998-f285-4cb4-e870-8e1e9efb5a1b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.76      0.17      0.28       239\n",
            "           1       0.26      0.89      0.40        81\n",
            "\n",
            "   micro avg       0.34      0.35      0.35       320\n",
            "   macro avg       0.51      0.53      0.34       320\n",
            "weighted avg       0.63      0.35      0.31       320\n",
            " samples avg       0.32      0.35      0.33       320\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##CNN ( 4 Conv layers + 3 dense layers )"
      ],
      "metadata": {
        "id": "lT_OJG2RDbPH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.layers import Input\n",
        "from keras.layers.merge import concatenate\n",
        "from keras.utils.vis_utils import plot_model\n",
        "\n",
        "\n",
        "intput_shape=(40, 8064) #x_train.shape[1]\n",
        "\n",
        "input_layer = Input(intput_shape)\n",
        "\n",
        "\n",
        "conv1 = Conv1D(32, kernel_size=6,padding = 'same',activation='relu') (input_layer)\n",
        "bn1 = BatchNormalization() (conv1)\n",
        "pool1 = MaxPooling1D(pool_size=(2)) (bn1)\n",
        "\n",
        "conv2 = Conv1D(32,kernel_size=6,padding = 'same', activation='relu') (pool1)\n",
        "bn2 = BatchNormalization() (conv2) \n",
        "pool2 = MaxPooling1D(pool_size=(2)) (bn2)\n",
        "\n",
        "conv3 = Conv1D(32,kernel_size=6,padding = 'same', activation='relu') (pool2)\n",
        "bn3 = BatchNormalization() (conv3)\n",
        "pool3 = MaxPooling1D(pool_size=(2)) (bn3)\n",
        "\n",
        "connect1 = concatenate([pool1, pool2, pool3], axis=1)\n",
        "\n",
        "conv4 = Conv1D(128,kernel_size=6,padding = 'same', activation='relu') (connect1)\n",
        "bn4 = BatchNormalization() (conv4)\n",
        "pool4 = MaxPooling1D(pool_size=(2)) (bn4)\n",
        "\n",
        "flat = Flatten() (pool4)\n",
        "\n",
        "dense1 = Dense(1024, activation='relu') (flat)\n",
        "drop1 = Dropout(0.2) (dense1)\n",
        "\n",
        "dense2 = Dense(256, activation='relu') (drop1)\n",
        "drop2 = Dropout(0.2) (dense2)\n",
        "\n",
        "dense3 = Dense(64, activation='relu') (drop2)\n",
        "drop3 = Dropout(0.2) (dense3)\n",
        "\n",
        "output_layer = Dense(2, activation='softmax') (drop3) #num_classes\n",
        "\n",
        "model = Model(input_layer, output_layer)\n",
        "model.summary()\n",
        "plot_model(model, to_file='model_plot.png', show_shapes=True, show_layer_names=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "lC75SU1BN5vm",
        "outputId": "324b9c03-d51f-4f4b-d065-7a98539dd84b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_2 (InputLayer)           [(None, 40, 8064)]   0           []                               \n",
            "                                                                                                  \n",
            " conv1d_4 (Conv1D)              (None, 40, 32)       1548320     ['input_2[0][0]']                \n",
            "                                                                                                  \n",
            " batch_normalization_4 (BatchNo  (None, 40, 32)      128         ['conv1d_4[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " max_pooling1d_4 (MaxPooling1D)  (None, 20, 32)      0           ['batch_normalization_4[0][0]']  \n",
            "                                                                                                  \n",
            " conv1d_5 (Conv1D)              (None, 20, 32)       6176        ['max_pooling1d_4[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_5 (BatchNo  (None, 20, 32)      128         ['conv1d_5[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " max_pooling1d_5 (MaxPooling1D)  (None, 10, 32)      0           ['batch_normalization_5[0][0]']  \n",
            "                                                                                                  \n",
            " conv1d_6 (Conv1D)              (None, 10, 32)       6176        ['max_pooling1d_5[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_6 (BatchNo  (None, 10, 32)      128         ['conv1d_6[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " max_pooling1d_6 (MaxPooling1D)  (None, 5, 32)       0           ['batch_normalization_6[0][0]']  \n",
            "                                                                                                  \n",
            " concatenate_1 (Concatenate)    (None, 35, 32)       0           ['max_pooling1d_4[0][0]',        \n",
            "                                                                  'max_pooling1d_5[0][0]',        \n",
            "                                                                  'max_pooling1d_6[0][0]']        \n",
            "                                                                                                  \n",
            " conv1d_7 (Conv1D)              (None, 35, 128)      24704       ['concatenate_1[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_7 (BatchNo  (None, 35, 128)     512         ['conv1d_7[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " max_pooling1d_7 (MaxPooling1D)  (None, 17, 128)     0           ['batch_normalization_7[0][0]']  \n",
            "                                                                                                  \n",
            " flatten_1 (Flatten)            (None, 2176)         0           ['max_pooling1d_7[0][0]']        \n",
            "                                                                                                  \n",
            " dense_4 (Dense)                (None, 1024)         2229248     ['flatten_1[0][0]']              \n",
            "                                                                                                  \n",
            " dropout_3 (Dropout)            (None, 1024)         0           ['dense_4[0][0]']                \n",
            "                                                                                                  \n",
            " dense_5 (Dense)                (None, 256)          262400      ['dropout_3[0][0]']              \n",
            "                                                                                                  \n",
            " dropout_4 (Dropout)            (None, 256)          0           ['dense_5[0][0]']                \n",
            "                                                                                                  \n",
            " dense_6 (Dense)                (None, 64)           16448       ['dropout_4[0][0]']              \n",
            "                                                                                                  \n",
            " dropout_5 (Dropout)            (None, 64)           0           ['dense_6[0][0]']                \n",
            "                                                                                                  \n",
            " dense_7 (Dense)                (None, 2)            130         ['dropout_5[0][0]']              \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 4,094,498\n",
            "Trainable params: 4,094,050\n",
            "Non-trainable params: 448\n",
            "__________________________________________________________________________________________________\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(loss=keras.losses.categorical_crossentropy, \n",
        "              optimizer='adam',\n",
        "              metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "zBtYqOvGGOfV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Arousal"
      ],
      "metadata": {
        "id": "q4TAMKt3rD1P"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# check points and early stopping\n",
        "from keras.callbacks import ModelCheckpoint,EarlyStopping\n",
        "\n",
        "model_name = \"\"\n",
        "filepath=\"/content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/\"  + model_name + \"weights-improvement-{epoch:02d}-{accuracy:.4f}.hdf5\"\n",
        "print(filepath)\n",
        "checkpoint = ModelCheckpoint(filepath, monitor='accuracy', verbose=1, save_best_only=True, mode='max')\n",
        "es = EarlyStopping(monitor='accuracy', mode='max', verbose=1, patience=10)\n",
        "callbacks_list = [es, checkpoint]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q0pefFUxrFXP",
        "outputId": "3f0f266d-7b36-4665-b74a-4fcec49cf2f6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-{epoch:02d}-{accuracy:.4f}.hdf5\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Raw M1 Arousal"
      ],
      "metadata": {
        "id": "Lv6S2bqBOY4d"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#validation_data=(x_val, y_val),\n",
        "H = model.fit(data_training, y_train,\n",
        "          batch_size=200,\n",
        "          epochs=100,\n",
        "          verbose=1,\n",
        "          validation_data=(data_testing, y_test), \n",
        "          callbacks= callbacks_list)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fkecuZOerI32",
        "outputId": "83dd8581-e31d-4b6d-8e37-5402ce052ef6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 1.1138 - accuracy: 0.5208\n",
            "Epoch 1: accuracy improved from -inf to 0.52083, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-01-0.5208.hdf5\n",
            "5/5 [==============================] - 15s 422ms/step - loss: 1.1138 - accuracy: 0.5208 - val_loss: 0.9506 - val_accuracy: 0.3969\n",
            "Epoch 2/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.9197 - accuracy: 0.5292\n",
            "Epoch 2: accuracy improved from 0.52083 to 0.52917, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-02-0.5292.hdf5\n",
            "5/5 [==============================] - 1s 202ms/step - loss: 0.9197 - accuracy: 0.5292 - val_loss: 0.7609 - val_accuracy: 0.6375\n",
            "Epoch 3/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.8681 - accuracy: 0.5354\n",
            "Epoch 3: accuracy improved from 0.52917 to 0.53542, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-03-0.5354.hdf5\n",
            "5/5 [==============================] - 1s 187ms/step - loss: 0.8681 - accuracy: 0.5354 - val_loss: 0.9095 - val_accuracy: 0.3438\n",
            "Epoch 4/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.8613 - accuracy: 0.5500\n",
            "Epoch 4: accuracy improved from 0.53542 to 0.55000, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-04-0.5500.hdf5\n",
            "5/5 [==============================] - 1s 196ms/step - loss: 0.8613 - accuracy: 0.5500 - val_loss: 0.9776 - val_accuracy: 0.3500\n",
            "Epoch 5/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7215 - accuracy: 0.5250\n",
            "Epoch 5: accuracy did not improve from 0.55000\n",
            "5/5 [==============================] - 1s 127ms/step - loss: 0.7215 - accuracy: 0.5250 - val_loss: 0.7307 - val_accuracy: 0.4625\n",
            "Epoch 6/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7351 - accuracy: 0.5594\n",
            "Epoch 6: accuracy improved from 0.55000 to 0.55937, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-06-0.5594.hdf5\n",
            "5/5 [==============================] - 1s 187ms/step - loss: 0.7351 - accuracy: 0.5594 - val_loss: 0.7251 - val_accuracy: 0.4437\n",
            "Epoch 7/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7051 - accuracy: 0.5406\n",
            "Epoch 7: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 134ms/step - loss: 0.7051 - accuracy: 0.5406 - val_loss: 0.7826 - val_accuracy: 0.3625\n",
            "Epoch 8/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7491 - accuracy: 0.5365\n",
            "Epoch 8: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 132ms/step - loss: 0.7491 - accuracy: 0.5365 - val_loss: 0.6858 - val_accuracy: 0.5312\n",
            "Epoch 9/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7306 - accuracy: 0.5354\n",
            "Epoch 9: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 132ms/step - loss: 0.7306 - accuracy: 0.5354 - val_loss: 0.6823 - val_accuracy: 0.7250\n",
            "Epoch 10/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7233 - accuracy: 0.5167\n",
            "Epoch 10: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 140ms/step - loss: 0.7233 - accuracy: 0.5167 - val_loss: 0.7104 - val_accuracy: 0.3938\n",
            "Epoch 11/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7254 - accuracy: 0.5573\n",
            "Epoch 11: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 131ms/step - loss: 0.7254 - accuracy: 0.5573 - val_loss: 0.7150 - val_accuracy: 0.4875\n",
            "Epoch 12/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7076 - accuracy: 0.5531\n",
            "Epoch 12: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 137ms/step - loss: 0.7076 - accuracy: 0.5531 - val_loss: 0.6733 - val_accuracy: 0.7219\n",
            "Epoch 13/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7116 - accuracy: 0.5490\n",
            "Epoch 13: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 131ms/step - loss: 0.7116 - accuracy: 0.5490 - val_loss: 0.6905 - val_accuracy: 0.6656\n",
            "Epoch 14/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6975 - accuracy: 0.5406\n",
            "Epoch 14: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.6975 - accuracy: 0.5406 - val_loss: 0.7110 - val_accuracy: 0.4938\n",
            "Epoch 15/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7061 - accuracy: 0.5594\n",
            "Epoch 15: accuracy did not improve from 0.55937\n",
            "5/5 [==============================] - 1s 132ms/step - loss: 0.7061 - accuracy: 0.5594 - val_loss: 0.6921 - val_accuracy: 0.5844\n",
            "Epoch 16/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7087 - accuracy: 0.5625\n",
            "Epoch 16: accuracy improved from 0.55937 to 0.56250, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-16-0.5625.hdf5\n",
            "5/5 [==============================] - 1s 248ms/step - loss: 0.7087 - accuracy: 0.5625 - val_loss: 0.6814 - val_accuracy: 0.6938\n",
            "Epoch 17/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7056 - accuracy: 0.5240\n",
            "Epoch 17: accuracy did not improve from 0.56250\n",
            "5/5 [==============================] - 1s 148ms/step - loss: 0.7056 - accuracy: 0.5240 - val_loss: 0.6816 - val_accuracy: 0.6656\n",
            "Epoch 18/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6934 - accuracy: 0.5604\n",
            "Epoch 18: accuracy did not improve from 0.56250\n",
            "5/5 [==============================] - 1s 138ms/step - loss: 0.6934 - accuracy: 0.5604 - val_loss: 0.6787 - val_accuracy: 0.7031\n",
            "Epoch 19/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6839 - accuracy: 0.5552\n",
            "Epoch 19: accuracy did not improve from 0.56250\n",
            "5/5 [==============================] - 1s 132ms/step - loss: 0.6839 - accuracy: 0.5552 - val_loss: 0.6795 - val_accuracy: 0.7063\n",
            "Epoch 20/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6830 - accuracy: 0.5625\n",
            "Epoch 20: accuracy did not improve from 0.56250\n",
            "5/5 [==============================] - 1s 134ms/step - loss: 0.6830 - accuracy: 0.5625 - val_loss: 0.6812 - val_accuracy: 0.6281\n",
            "Epoch 21/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6780 - accuracy: 0.5594\n",
            "Epoch 21: accuracy did not improve from 0.56250\n",
            "5/5 [==============================] - 1s 132ms/step - loss: 0.6780 - accuracy: 0.5594 - val_loss: 0.6808 - val_accuracy: 0.6125\n",
            "Epoch 22/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6844 - accuracy: 0.5813\n",
            "Epoch 22: accuracy improved from 0.56250 to 0.58125, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-22-0.5813.hdf5\n",
            "5/5 [==============================] - 1s 233ms/step - loss: 0.6844 - accuracy: 0.5813 - val_loss: 0.6738 - val_accuracy: 0.7250\n",
            "Epoch 23/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6706 - accuracy: 0.5750\n",
            "Epoch 23: accuracy did not improve from 0.58125\n",
            "5/5 [==============================] - 1s 130ms/step - loss: 0.6706 - accuracy: 0.5750 - val_loss: 0.6752 - val_accuracy: 0.6750\n",
            "Epoch 24/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6674 - accuracy: 0.5938\n",
            "Epoch 24: accuracy improved from 0.58125 to 0.59375, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-24-0.5938.hdf5\n",
            "5/5 [==============================] - 1s 197ms/step - loss: 0.6674 - accuracy: 0.5938 - val_loss: 0.6776 - val_accuracy: 0.6000\n",
            "Epoch 25/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6869 - accuracy: 0.6000\n",
            "Epoch 25: accuracy improved from 0.59375 to 0.60000, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-25-0.6000.hdf5\n",
            "5/5 [==============================] - 1s 194ms/step - loss: 0.6869 - accuracy: 0.6000 - val_loss: 0.6799 - val_accuracy: 0.5437\n",
            "Epoch 26/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6716 - accuracy: 0.5854\n",
            "Epoch 26: accuracy did not improve from 0.60000\n",
            "5/5 [==============================] - 1s 133ms/step - loss: 0.6716 - accuracy: 0.5854 - val_loss: 0.6715 - val_accuracy: 0.6219\n",
            "Epoch 27/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6602 - accuracy: 0.6031\n",
            "Epoch 27: accuracy improved from 0.60000 to 0.60312, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-27-0.6031.hdf5\n",
            "5/5 [==============================] - 1s 194ms/step - loss: 0.6602 - accuracy: 0.6031 - val_loss: 0.6784 - val_accuracy: 0.5500\n",
            "Epoch 28/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6541 - accuracy: 0.6187\n",
            "Epoch 28: accuracy improved from 0.60312 to 0.61875, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-28-0.6187.hdf5\n",
            "5/5 [==============================] - 1s 192ms/step - loss: 0.6541 - accuracy: 0.6187 - val_loss: 0.6729 - val_accuracy: 0.6125\n",
            "Epoch 29/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6479 - accuracy: 0.6271\n",
            "Epoch 29: accuracy improved from 0.61875 to 0.62708, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-29-0.6271.hdf5\n",
            "5/5 [==============================] - 1s 210ms/step - loss: 0.6479 - accuracy: 0.6271 - val_loss: 0.6722 - val_accuracy: 0.5875\n",
            "Epoch 30/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6445 - accuracy: 0.6250\n",
            "Epoch 30: accuracy did not improve from 0.62708\n",
            "5/5 [==============================] - 1s 156ms/step - loss: 0.6445 - accuracy: 0.6250 - val_loss: 0.6587 - val_accuracy: 0.6625\n",
            "Epoch 31/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6514 - accuracy: 0.6115\n",
            "Epoch 31: accuracy did not improve from 0.62708\n",
            "5/5 [==============================] - 1s 136ms/step - loss: 0.6514 - accuracy: 0.6115 - val_loss: 0.6555 - val_accuracy: 0.7094\n",
            "Epoch 32/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6296 - accuracy: 0.6198\n",
            "Epoch 32: accuracy did not improve from 0.62708\n",
            "5/5 [==============================] - 1s 134ms/step - loss: 0.6296 - accuracy: 0.6198 - val_loss: 0.6577 - val_accuracy: 0.6625\n",
            "Epoch 33/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6204 - accuracy: 0.6490\n",
            "Epoch 33: accuracy improved from 0.62708 to 0.64896, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-33-0.6490.hdf5\n",
            "5/5 [==============================] - 1s 227ms/step - loss: 0.6204 - accuracy: 0.6490 - val_loss: 0.6308 - val_accuracy: 0.7219\n",
            "Epoch 34/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6206 - accuracy: 0.6313\n",
            "Epoch 34: accuracy did not improve from 0.64896\n",
            "5/5 [==============================] - 1s 138ms/step - loss: 0.6206 - accuracy: 0.6313 - val_loss: 0.6689 - val_accuracy: 0.5938\n",
            "Epoch 35/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5892 - accuracy: 0.6677\n",
            "Epoch 35: accuracy improved from 0.64896 to 0.66771, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-35-0.6677.hdf5\n",
            "5/5 [==============================] - 1s 211ms/step - loss: 0.5892 - accuracy: 0.6677 - val_loss: 0.6547 - val_accuracy: 0.6844\n",
            "Epoch 36/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5835 - accuracy: 0.6667\n",
            "Epoch 36: accuracy did not improve from 0.66771\n",
            "5/5 [==============================] - 1s 158ms/step - loss: 0.5835 - accuracy: 0.6667 - val_loss: 0.6804 - val_accuracy: 0.6687\n",
            "Epoch 37/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5899 - accuracy: 0.6812\n",
            "Epoch 37: accuracy improved from 0.66771 to 0.68125, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-37-0.6812.hdf5\n",
            "5/5 [==============================] - 1s 229ms/step - loss: 0.5899 - accuracy: 0.6812 - val_loss: 0.6718 - val_accuracy: 0.6594\n",
            "Epoch 38/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5761 - accuracy: 0.6917\n",
            "Epoch 38: accuracy improved from 0.68125 to 0.69167, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-38-0.6917.hdf5\n",
            "5/5 [==============================] - 1s 234ms/step - loss: 0.5761 - accuracy: 0.6917 - val_loss: 0.6573 - val_accuracy: 0.6656\n",
            "Epoch 39/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5624 - accuracy: 0.6844\n",
            "Epoch 39: accuracy did not improve from 0.69167\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.5624 - accuracy: 0.6844 - val_loss: 0.6361 - val_accuracy: 0.7281\n",
            "Epoch 40/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5101 - accuracy: 0.7229\n",
            "Epoch 40: accuracy improved from 0.69167 to 0.72292, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-40-0.7229.hdf5\n",
            "5/5 [==============================] - 1s 188ms/step - loss: 0.5101 - accuracy: 0.7229 - val_loss: 0.6573 - val_accuracy: 0.7094\n",
            "Epoch 41/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5019 - accuracy: 0.7281\n",
            "Epoch 41: accuracy improved from 0.72292 to 0.72812, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-41-0.7281.hdf5\n",
            "5/5 [==============================] - 1s 197ms/step - loss: 0.5019 - accuracy: 0.7281 - val_loss: 0.6564 - val_accuracy: 0.7000\n",
            "Epoch 42/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5016 - accuracy: 0.7302\n",
            "Epoch 42: accuracy improved from 0.72812 to 0.73021, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-42-0.7302.hdf5\n",
            "5/5 [==============================] - 1s 199ms/step - loss: 0.5016 - accuracy: 0.7302 - val_loss: 0.6454 - val_accuracy: 0.6906\n",
            "Epoch 43/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5270 - accuracy: 0.7094\n",
            "Epoch 43: accuracy did not improve from 0.73021\n",
            "5/5 [==============================] - 1s 151ms/step - loss: 0.5270 - accuracy: 0.7094 - val_loss: 0.6736 - val_accuracy: 0.6750\n",
            "Epoch 44/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5406 - accuracy: 0.6958\n",
            "Epoch 44: accuracy did not improve from 0.73021\n",
            "5/5 [==============================] - 1s 128ms/step - loss: 0.5406 - accuracy: 0.6958 - val_loss: 0.6980 - val_accuracy: 0.6375\n",
            "Epoch 45/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5536 - accuracy: 0.6885\n",
            "Epoch 45: accuracy did not improve from 0.73021\n",
            "5/5 [==============================] - 1s 151ms/step - loss: 0.5536 - accuracy: 0.6885 - val_loss: 0.6890 - val_accuracy: 0.6406\n",
            "Epoch 46/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5629 - accuracy: 0.6844\n",
            "Epoch 46: accuracy did not improve from 0.73021\n",
            "5/5 [==============================] - 1s 158ms/step - loss: 0.5629 - accuracy: 0.6844 - val_loss: 0.6950 - val_accuracy: 0.7188\n",
            "Epoch 47/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5198 - accuracy: 0.7031\n",
            "Epoch 47: accuracy did not improve from 0.73021\n",
            "5/5 [==============================] - 1s 150ms/step - loss: 0.5198 - accuracy: 0.7031 - val_loss: 0.6475 - val_accuracy: 0.7281\n",
            "Epoch 48/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4712 - accuracy: 0.7458\n",
            "Epoch 48: accuracy improved from 0.73021 to 0.74583, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-48-0.7458.hdf5\n",
            "5/5 [==============================] - 1s 263ms/step - loss: 0.4712 - accuracy: 0.7458 - val_loss: 0.6691 - val_accuracy: 0.7281\n",
            "Epoch 49/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4513 - accuracy: 0.7708\n",
            "Epoch 49: accuracy improved from 0.74583 to 0.77083, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-49-0.7708.hdf5\n",
            "5/5 [==============================] - 1s 251ms/step - loss: 0.4513 - accuracy: 0.7708 - val_loss: 0.6451 - val_accuracy: 0.7219\n",
            "Epoch 50/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4667 - accuracy: 0.7479\n",
            "Epoch 50: accuracy did not improve from 0.77083\n",
            "5/5 [==============================] - 1s 150ms/step - loss: 0.4667 - accuracy: 0.7479 - val_loss: 0.6646 - val_accuracy: 0.7156\n",
            "Epoch 51/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5043 - accuracy: 0.7542\n",
            "Epoch 51: accuracy did not improve from 0.77083\n",
            "5/5 [==============================] - 1s 137ms/step - loss: 0.5043 - accuracy: 0.7542 - val_loss: 0.6624 - val_accuracy: 0.7281\n",
            "Epoch 52/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4565 - accuracy: 0.7479\n",
            "Epoch 52: accuracy did not improve from 0.77083\n",
            "5/5 [==============================] - 1s 126ms/step - loss: 0.4565 - accuracy: 0.7479 - val_loss: 0.6893 - val_accuracy: 0.7219\n",
            "Epoch 53/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4394 - accuracy: 0.7646\n",
            "Epoch 53: accuracy did not improve from 0.77083\n",
            "5/5 [==============================] - 1s 130ms/step - loss: 0.4394 - accuracy: 0.7646 - val_loss: 0.7163 - val_accuracy: 0.6594\n",
            "Epoch 54/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4684 - accuracy: 0.7740\n",
            "Epoch 54: accuracy improved from 0.77083 to 0.77396, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-54-0.7740.hdf5\n",
            "5/5 [==============================] - 1s 192ms/step - loss: 0.4684 - accuracy: 0.7740 - val_loss: 0.7177 - val_accuracy: 0.7250\n",
            "Epoch 55/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4623 - accuracy: 0.7542\n",
            "Epoch 55: accuracy did not improve from 0.77396\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.4623 - accuracy: 0.7542 - val_loss: 0.7424 - val_accuracy: 0.7188\n",
            "Epoch 56/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4475 - accuracy: 0.7656\n",
            "Epoch 56: accuracy did not improve from 0.77396\n",
            "5/5 [==============================] - 1s 133ms/step - loss: 0.4475 - accuracy: 0.7656 - val_loss: 0.7335 - val_accuracy: 0.7219\n",
            "Epoch 57/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4578 - accuracy: 0.7740\n",
            "Epoch 57: accuracy did not improve from 0.77396\n",
            "5/5 [==============================] - 1s 134ms/step - loss: 0.4578 - accuracy: 0.7740 - val_loss: 0.7145 - val_accuracy: 0.7063\n",
            "Epoch 58/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4089 - accuracy: 0.7896\n",
            "Epoch 58: accuracy improved from 0.77396 to 0.78958, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-58-0.7896.hdf5\n",
            "5/5 [==============================] - 1s 192ms/step - loss: 0.4089 - accuracy: 0.7896 - val_loss: 0.7646 - val_accuracy: 0.7219\n",
            "Epoch 59/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4096 - accuracy: 0.8062\n",
            "Epoch 59: accuracy improved from 0.78958 to 0.80625, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-59-0.8062.hdf5\n",
            "5/5 [==============================] - 1s 192ms/step - loss: 0.4096 - accuracy: 0.8062 - val_loss: 0.7284 - val_accuracy: 0.7281\n",
            "Epoch 60/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3908 - accuracy: 0.8062\n",
            "Epoch 60: accuracy did not improve from 0.80625\n",
            "5/5 [==============================] - 1s 142ms/step - loss: 0.3908 - accuracy: 0.8062 - val_loss: 0.7529 - val_accuracy: 0.7031\n",
            "Epoch 61/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3967 - accuracy: 0.7802\n",
            "Epoch 61: accuracy did not improve from 0.80625\n",
            "5/5 [==============================] - 1s 143ms/step - loss: 0.3967 - accuracy: 0.7802 - val_loss: 0.7316 - val_accuracy: 0.6844\n",
            "Epoch 62/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3679 - accuracy: 0.8292\n",
            "Epoch 62: accuracy improved from 0.80625 to 0.82917, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-62-0.8292.hdf5\n",
            "5/5 [==============================] - 1s 234ms/step - loss: 0.3679 - accuracy: 0.8292 - val_loss: 0.7391 - val_accuracy: 0.6906\n",
            "Epoch 63/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3145 - accuracy: 0.8521\n",
            "Epoch 63: accuracy improved from 0.82917 to 0.85208, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-63-0.8521.hdf5\n",
            "5/5 [==============================] - 1s 194ms/step - loss: 0.3145 - accuracy: 0.8521 - val_loss: 0.8408 - val_accuracy: 0.6875\n",
            "Epoch 64/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3007 - accuracy: 0.8667\n",
            "Epoch 64: accuracy improved from 0.85208 to 0.86667, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Arousal/Raw_MDPI_M1/weights-improvement-64-0.8667.hdf5\n",
            "5/5 [==============================] - 1s 191ms/step - loss: 0.3007 - accuracy: 0.8667 - val_loss: 0.8125 - val_accuracy: 0.6875\n",
            "Epoch 65/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4479 - accuracy: 0.7896\n",
            "Epoch 65: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 150ms/step - loss: 0.4479 - accuracy: 0.7896 - val_loss: 0.7404 - val_accuracy: 0.7188\n",
            "Epoch 66/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4597 - accuracy: 0.7688\n",
            "Epoch 66: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 134ms/step - loss: 0.4597 - accuracy: 0.7688 - val_loss: 0.6871 - val_accuracy: 0.7063\n",
            "Epoch 67/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4154 - accuracy: 0.7917\n",
            "Epoch 67: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 132ms/step - loss: 0.4154 - accuracy: 0.7917 - val_loss: 0.7229 - val_accuracy: 0.7094\n",
            "Epoch 68/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4153 - accuracy: 0.7875\n",
            "Epoch 68: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 138ms/step - loss: 0.4153 - accuracy: 0.7875 - val_loss: 0.7123 - val_accuracy: 0.6875\n",
            "Epoch 69/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4242 - accuracy: 0.7927\n",
            "Epoch 69: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 142ms/step - loss: 0.4242 - accuracy: 0.7927 - val_loss: 0.7153 - val_accuracy: 0.6781\n",
            "Epoch 70/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5198 - accuracy: 0.7302\n",
            "Epoch 70: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 131ms/step - loss: 0.5198 - accuracy: 0.7302 - val_loss: 0.7079 - val_accuracy: 0.6531\n",
            "Epoch 71/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5090 - accuracy: 0.7271\n",
            "Epoch 71: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 133ms/step - loss: 0.5090 - accuracy: 0.7271 - val_loss: 0.7221 - val_accuracy: 0.6281\n",
            "Epoch 72/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4562 - accuracy: 0.7708\n",
            "Epoch 72: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 154ms/step - loss: 0.4562 - accuracy: 0.7708 - val_loss: 0.7266 - val_accuracy: 0.6187\n",
            "Epoch 73/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4439 - accuracy: 0.7760\n",
            "Epoch 73: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 136ms/step - loss: 0.4439 - accuracy: 0.7760 - val_loss: 0.7640 - val_accuracy: 0.5781\n",
            "Epoch 74/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4109 - accuracy: 0.7969\n",
            "Epoch 74: accuracy did not improve from 0.86667\n",
            "5/5 [==============================] - 1s 141ms/step - loss: 0.4109 - accuracy: 0.7969 - val_loss: 0.8562 - val_accuracy: 0.6625\n",
            "Epoch 74: early stopping\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Valance"
      ],
      "metadata": {
        "id": "Qm1cbmmUq6Zp"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Automatic save \n",
        "(We have saved all the details of every epoch in google drive so that we can use this when we need.)**\n"
      ],
      "metadata": {
        "id": "ttoQN3bBGoTT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# check points and early stopping\n",
        "from keras.callbacks import ModelCheckpoint,EarlyStopping\n",
        "\n",
        "model_name = \"\"\n",
        "filepath=\"/content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/\"  + model_name + \"weights-improvement-{epoch:02d}-{accuracy:.4f}.hdf5\"\n",
        "print(filepath)\n",
        "checkpoint = ModelCheckpoint(filepath, monitor='accuracy', verbose=1, save_best_only=True, mode='max')\n",
        "es = EarlyStopping(monitor='accuracy', mode='max', verbose=1, patience=10)\n",
        "callbacks_list = [es, checkpoint]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "92821b77-de92-4292-b5ce-c845cc36f211",
        "id": "XR0N9xPxemmM"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-{epoch:02d}-{accuracy:.4f}.hdf5\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Fit the model M1 Valence**\n"
      ],
      "metadata": {
        "id": "Mduy9w8SHG3h"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#validation_data=(x_val, y_val),\n",
        "H = model.fit(data_training, y_train,\n",
        "          batch_size=200,\n",
        "          epochs=100,\n",
        "          verbose=1,\n",
        "          validation_data=(data_testing, y_test), \n",
        "          callbacks= callbacks_list)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RIwYlux0jotk",
        "outputId": "82008606-a102-47b7-a6a3-a5e592458a89"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.8771 - accuracy: 0.6240\n",
            "Epoch 1: accuracy improved from -inf to 0.62396, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-01-0.6240.hdf5\n",
            "5/5 [==============================] - 4s 372ms/step - loss: 0.8771 - accuracy: 0.6240 - val_loss: 2.6886 - val_accuracy: 0.2531\n",
            "Epoch 2/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.8479 - accuracy: 0.6260\n",
            "Epoch 2: accuracy improved from 0.62396 to 0.62604, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-02-0.6260.hdf5\n",
            "5/5 [==============================] - 1s 208ms/step - loss: 0.8479 - accuracy: 0.6260 - val_loss: 1.7303 - val_accuracy: 0.2531\n",
            "Epoch 3/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.8672 - accuracy: 0.5865\n",
            "Epoch 3: accuracy did not improve from 0.62604\n",
            "5/5 [==============================] - 1s 139ms/step - loss: 0.8672 - accuracy: 0.5865 - val_loss: 1.2556 - val_accuracy: 0.2531\n",
            "Epoch 4/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.8218 - accuracy: 0.6667\n",
            "Epoch 4: accuracy improved from 0.62604 to 0.66667, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-04-0.6667.hdf5\n",
            "5/5 [==============================] - 1s 201ms/step - loss: 0.8218 - accuracy: 0.6667 - val_loss: 1.2763 - val_accuracy: 0.2531\n",
            "Epoch 5/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6767 - accuracy: 0.6458\n",
            "Epoch 5: accuracy did not improve from 0.66667\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.6767 - accuracy: 0.6458 - val_loss: 0.8716 - val_accuracy: 0.6156\n",
            "Epoch 6/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.7565 - accuracy: 0.6396\n",
            "Epoch 6: accuracy did not improve from 0.66667\n",
            "5/5 [==============================] - 1s 151ms/step - loss: 0.7565 - accuracy: 0.6396 - val_loss: 0.9082 - val_accuracy: 0.2531\n",
            "Epoch 7/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6622 - accuracy: 0.6542\n",
            "Epoch 7: accuracy did not improve from 0.66667\n",
            "5/5 [==============================] - 1s 155ms/step - loss: 0.6622 - accuracy: 0.6542 - val_loss: 0.7389 - val_accuracy: 0.4062\n",
            "Epoch 8/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6639 - accuracy: 0.6667\n",
            "Epoch 8: accuracy did not improve from 0.66667\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.6639 - accuracy: 0.6667 - val_loss: 0.7525 - val_accuracy: 0.2594\n",
            "Epoch 9/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6864 - accuracy: 0.6583\n",
            "Epoch 9: accuracy did not improve from 0.66667\n",
            "5/5 [==============================] - 1s 134ms/step - loss: 0.6864 - accuracy: 0.6583 - val_loss: 0.7281 - val_accuracy: 0.4938\n",
            "Epoch 10/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6600 - accuracy: 0.6635\n",
            "Epoch 10: accuracy did not improve from 0.66667\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.6600 - accuracy: 0.6635 - val_loss: 0.7123 - val_accuracy: 0.4031\n",
            "Epoch 11/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6579 - accuracy: 0.6729\n",
            "Epoch 11: accuracy improved from 0.66667 to 0.67292, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-11-0.6729.hdf5\n",
            "5/5 [==============================] - 1s 196ms/step - loss: 0.6579 - accuracy: 0.6729 - val_loss: 0.7777 - val_accuracy: 0.2531\n",
            "Epoch 12/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6634 - accuracy: 0.6667\n",
            "Epoch 12: accuracy did not improve from 0.67292\n",
            "5/5 [==============================] - 1s 137ms/step - loss: 0.6634 - accuracy: 0.6667 - val_loss: 0.7300 - val_accuracy: 0.3031\n",
            "Epoch 13/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6396 - accuracy: 0.6792\n",
            "Epoch 13: accuracy improved from 0.67292 to 0.67917, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-13-0.6792.hdf5\n",
            "5/5 [==============================] - 1s 211ms/step - loss: 0.6396 - accuracy: 0.6792 - val_loss: 0.7118 - val_accuracy: 0.3594\n",
            "Epoch 14/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6459 - accuracy: 0.6656\n",
            "Epoch 14: accuracy did not improve from 0.67917\n",
            "5/5 [==============================] - 1s 140ms/step - loss: 0.6459 - accuracy: 0.6656 - val_loss: 0.7187 - val_accuracy: 0.3594\n",
            "Epoch 15/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6255 - accuracy: 0.6604\n",
            "Epoch 15: accuracy did not improve from 0.67917\n",
            "5/5 [==============================] - 1s 150ms/step - loss: 0.6255 - accuracy: 0.6604 - val_loss: 0.7325 - val_accuracy: 0.3125\n",
            "Epoch 16/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6239 - accuracy: 0.6667\n",
            "Epoch 16: accuracy did not improve from 0.67917\n",
            "5/5 [==============================] - 1s 158ms/step - loss: 0.6239 - accuracy: 0.6667 - val_loss: 0.7422 - val_accuracy: 0.2906\n",
            "Epoch 17/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6327 - accuracy: 0.6812\n",
            "Epoch 17: accuracy improved from 0.67917 to 0.68125, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-17-0.6812.hdf5\n",
            "5/5 [==============================] - 1s 198ms/step - loss: 0.6327 - accuracy: 0.6812 - val_loss: 0.7439 - val_accuracy: 0.2688\n",
            "Epoch 18/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6234 - accuracy: 0.6687\n",
            "Epoch 18: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 139ms/step - loss: 0.6234 - accuracy: 0.6687 - val_loss: 0.7409 - val_accuracy: 0.2875\n",
            "Epoch 19/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6162 - accuracy: 0.6698\n",
            "Epoch 19: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 148ms/step - loss: 0.6162 - accuracy: 0.6698 - val_loss: 0.7516 - val_accuracy: 0.2937\n",
            "Epoch 20/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6083 - accuracy: 0.6781\n",
            "Epoch 20: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.6083 - accuracy: 0.6781 - val_loss: 0.7445 - val_accuracy: 0.2937\n",
            "Epoch 21/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.6003 - accuracy: 0.6677\n",
            "Epoch 21: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 146ms/step - loss: 0.6003 - accuracy: 0.6677 - val_loss: 0.7464 - val_accuracy: 0.2937\n",
            "Epoch 22/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5998 - accuracy: 0.6687\n",
            "Epoch 22: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 145ms/step - loss: 0.5998 - accuracy: 0.6687 - val_loss: 0.7604 - val_accuracy: 0.2562\n",
            "Epoch 23/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5941 - accuracy: 0.6687\n",
            "Epoch 23: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.5941 - accuracy: 0.6687 - val_loss: 0.7827 - val_accuracy: 0.2531\n",
            "Epoch 24/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5953 - accuracy: 0.6771\n",
            "Epoch 24: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 136ms/step - loss: 0.5953 - accuracy: 0.6771 - val_loss: 0.7742 - val_accuracy: 0.2531\n",
            "Epoch 25/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5889 - accuracy: 0.6812\n",
            "Epoch 25: accuracy did not improve from 0.68125\n",
            "5/5 [==============================] - 1s 132ms/step - loss: 0.5889 - accuracy: 0.6812 - val_loss: 0.7839 - val_accuracy: 0.2531\n",
            "Epoch 26/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5728 - accuracy: 0.6906\n",
            "Epoch 26: accuracy improved from 0.68125 to 0.69063, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-26-0.6906.hdf5\n",
            "5/5 [==============================] - 1s 197ms/step - loss: 0.5728 - accuracy: 0.6906 - val_loss: 0.7738 - val_accuracy: 0.2594\n",
            "Epoch 27/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5416 - accuracy: 0.6990\n",
            "Epoch 27: accuracy improved from 0.69063 to 0.69896, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-27-0.6990.hdf5\n",
            "5/5 [==============================] - 1s 209ms/step - loss: 0.5416 - accuracy: 0.6990 - val_loss: 0.8428 - val_accuracy: 0.2562\n",
            "Epoch 28/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5563 - accuracy: 0.6906\n",
            "Epoch 28: accuracy did not improve from 0.69896\n",
            "5/5 [==============================] - 1s 138ms/step - loss: 0.5563 - accuracy: 0.6906 - val_loss: 0.8672 - val_accuracy: 0.2656\n",
            "Epoch 29/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5726 - accuracy: 0.6865\n",
            "Epoch 29: accuracy did not improve from 0.69896\n",
            "5/5 [==============================] - 1s 200ms/step - loss: 0.5726 - accuracy: 0.6865 - val_loss: 0.7718 - val_accuracy: 0.2906\n",
            "Epoch 30/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5719 - accuracy: 0.6938\n",
            "Epoch 30: accuracy did not improve from 0.69896\n",
            "5/5 [==============================] - 1s 169ms/step - loss: 0.5719 - accuracy: 0.6938 - val_loss: 0.8954 - val_accuracy: 0.2750\n",
            "Epoch 31/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5520 - accuracy: 0.6792\n",
            "Epoch 31: accuracy did not improve from 0.69896\n",
            "5/5 [==============================] - 1s 175ms/step - loss: 0.5520 - accuracy: 0.6792 - val_loss: 0.7793 - val_accuracy: 0.2656\n",
            "Epoch 32/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5779 - accuracy: 0.7167\n",
            "Epoch 32: accuracy improved from 0.69896 to 0.71667, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-32-0.7167.hdf5\n",
            "5/5 [==============================] - 1s 238ms/step - loss: 0.5779 - accuracy: 0.7167 - val_loss: 0.8157 - val_accuracy: 0.2625\n",
            "Epoch 33/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5350 - accuracy: 0.7000\n",
            "Epoch 33: accuracy did not improve from 0.71667\n",
            "5/5 [==============================] - 1s 185ms/step - loss: 0.5350 - accuracy: 0.7000 - val_loss: 0.8523 - val_accuracy: 0.2906\n",
            "Epoch 34/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5510 - accuracy: 0.7156\n",
            "Epoch 34: accuracy did not improve from 0.71667\n",
            "5/5 [==============================] - 1s 170ms/step - loss: 0.5510 - accuracy: 0.7156 - val_loss: 0.7982 - val_accuracy: 0.3156\n",
            "Epoch 35/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5150 - accuracy: 0.7281\n",
            "Epoch 35: accuracy improved from 0.71667 to 0.72812, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-35-0.7281.hdf5\n",
            "5/5 [==============================] - 1s 313ms/step - loss: 0.5150 - accuracy: 0.7281 - val_loss: 0.9413 - val_accuracy: 0.2562\n",
            "Epoch 36/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5007 - accuracy: 0.7437\n",
            "Epoch 36: accuracy improved from 0.72812 to 0.74375, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-36-0.7437.hdf5\n",
            "5/5 [==============================] - 1s 272ms/step - loss: 0.5007 - accuracy: 0.7437 - val_loss: 0.8989 - val_accuracy: 0.2719\n",
            "Epoch 37/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.5059 - accuracy: 0.7406\n",
            "Epoch 37: accuracy did not improve from 0.74375\n",
            "5/5 [==============================] - 1s 162ms/step - loss: 0.5059 - accuracy: 0.7406 - val_loss: 0.9224 - val_accuracy: 0.3000\n",
            "Epoch 38/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4542 - accuracy: 0.7635\n",
            "Epoch 38: accuracy improved from 0.74375 to 0.76354, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-38-0.7635.hdf5\n",
            "5/5 [==============================] - 1s 285ms/step - loss: 0.4542 - accuracy: 0.7635 - val_loss: 1.0500 - val_accuracy: 0.3000\n",
            "Epoch 39/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4609 - accuracy: 0.7667\n",
            "Epoch 39: accuracy improved from 0.76354 to 0.76667, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-39-0.7667.hdf5\n",
            "5/5 [==============================] - 1s 220ms/step - loss: 0.4609 - accuracy: 0.7667 - val_loss: 0.9704 - val_accuracy: 0.2719\n",
            "Epoch 40/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4614 - accuracy: 0.7823\n",
            "Epoch 40: accuracy improved from 0.76667 to 0.78229, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-40-0.7823.hdf5\n",
            "5/5 [==============================] - 1s 197ms/step - loss: 0.4614 - accuracy: 0.7823 - val_loss: 1.1011 - val_accuracy: 0.2625\n",
            "Epoch 41/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4119 - accuracy: 0.7958\n",
            "Epoch 41: accuracy improved from 0.78229 to 0.79583, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-41-0.7958.hdf5\n",
            "5/5 [==============================] - 1s 210ms/step - loss: 0.4119 - accuracy: 0.7958 - val_loss: 0.9836 - val_accuracy: 0.2844\n",
            "Epoch 42/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3942 - accuracy: 0.8000\n",
            "Epoch 42: accuracy improved from 0.79583 to 0.80000, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-42-0.8000.hdf5\n",
            "5/5 [==============================] - 1s 195ms/step - loss: 0.3942 - accuracy: 0.8000 - val_loss: 1.0799 - val_accuracy: 0.2719\n",
            "Epoch 43/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3992 - accuracy: 0.7948\n",
            "Epoch 43: accuracy did not improve from 0.80000\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.3992 - accuracy: 0.7948 - val_loss: 1.0113 - val_accuracy: 0.3094\n",
            "Epoch 44/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3967 - accuracy: 0.7854\n",
            "Epoch 44: accuracy did not improve from 0.80000\n",
            "5/5 [==============================] - 1s 163ms/step - loss: 0.3967 - accuracy: 0.7854 - val_loss: 1.0292 - val_accuracy: 0.2781\n",
            "Epoch 45/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4259 - accuracy: 0.7937\n",
            "Epoch 45: accuracy did not improve from 0.80000\n",
            "5/5 [==============================] - 1s 172ms/step - loss: 0.4259 - accuracy: 0.7937 - val_loss: 1.1101 - val_accuracy: 0.2750\n",
            "Epoch 46/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4311 - accuracy: 0.7948\n",
            "Epoch 46: accuracy did not improve from 0.80000\n",
            "5/5 [==============================] - 1s 140ms/step - loss: 0.4311 - accuracy: 0.7948 - val_loss: 0.9343 - val_accuracy: 0.3313\n",
            "Epoch 47/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4364 - accuracy: 0.7563\n",
            "Epoch 47: accuracy did not improve from 0.80000\n",
            "5/5 [==============================] - 1s 163ms/step - loss: 0.4364 - accuracy: 0.7563 - val_loss: 1.0414 - val_accuracy: 0.3063\n",
            "Epoch 48/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3874 - accuracy: 0.8042\n",
            "Epoch 48: accuracy improved from 0.80000 to 0.80417, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-48-0.8042.hdf5\n",
            "5/5 [==============================] - 1s 207ms/step - loss: 0.3874 - accuracy: 0.8042 - val_loss: 1.1374 - val_accuracy: 0.2906\n",
            "Epoch 49/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3616 - accuracy: 0.8156\n",
            "Epoch 49: accuracy improved from 0.80417 to 0.81563, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-49-0.8156.hdf5\n",
            "5/5 [==============================] - 1s 296ms/step - loss: 0.3616 - accuracy: 0.8156 - val_loss: 1.3334 - val_accuracy: 0.3063\n",
            "Epoch 50/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4439 - accuracy: 0.7823\n",
            "Epoch 50: accuracy did not improve from 0.81563\n",
            "5/5 [==============================] - 1s 205ms/step - loss: 0.4439 - accuracy: 0.7823 - val_loss: 0.9327 - val_accuracy: 0.2750\n",
            "Epoch 51/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4699 - accuracy: 0.7917\n",
            "Epoch 51: accuracy did not improve from 0.81563\n",
            "5/5 [==============================] - 1s 154ms/step - loss: 0.4699 - accuracy: 0.7917 - val_loss: 1.1046 - val_accuracy: 0.3469\n",
            "Epoch 52/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.4574 - accuracy: 0.7667\n",
            "Epoch 52: accuracy did not improve from 0.81563\n",
            "5/5 [==============================] - 1s 147ms/step - loss: 0.4574 - accuracy: 0.7667 - val_loss: 1.0091 - val_accuracy: 0.3156\n",
            "Epoch 53/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3943 - accuracy: 0.8010\n",
            "Epoch 53: accuracy did not improve from 0.81563\n",
            "5/5 [==============================] - 1s 167ms/step - loss: 0.3943 - accuracy: 0.8010 - val_loss: 1.2079 - val_accuracy: 0.2781\n",
            "Epoch 54/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3461 - accuracy: 0.8135\n",
            "Epoch 54: accuracy did not improve from 0.81563\n",
            "5/5 [==============================] - 1s 139ms/step - loss: 0.3461 - accuracy: 0.8135 - val_loss: 1.3179 - val_accuracy: 0.2844\n",
            "Epoch 55/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3863 - accuracy: 0.8000\n",
            "Epoch 55: accuracy did not improve from 0.81563\n",
            "5/5 [==============================] - 1s 135ms/step - loss: 0.3863 - accuracy: 0.8000 - val_loss: 1.4460 - val_accuracy: 0.2844\n",
            "Epoch 56/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3440 - accuracy: 0.8188\n",
            "Epoch 56: accuracy improved from 0.81563 to 0.81875, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-56-0.8188.hdf5\n",
            "5/5 [==============================] - 1s 198ms/step - loss: 0.3440 - accuracy: 0.8188 - val_loss: 1.4238 - val_accuracy: 0.3125\n",
            "Epoch 57/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3674 - accuracy: 0.8021\n",
            "Epoch 57: accuracy did not improve from 0.81875\n",
            "5/5 [==============================] - 1s 141ms/step - loss: 0.3674 - accuracy: 0.8021 - val_loss: 1.4300 - val_accuracy: 0.3094\n",
            "Epoch 58/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3252 - accuracy: 0.8281\n",
            "Epoch 58: accuracy improved from 0.81875 to 0.82812, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-58-0.8281.hdf5\n",
            "5/5 [==============================] - 1s 237ms/step - loss: 0.3252 - accuracy: 0.8281 - val_loss: 1.6362 - val_accuracy: 0.2875\n",
            "Epoch 59/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3343 - accuracy: 0.8427\n",
            "Epoch 59: accuracy improved from 0.82812 to 0.84271, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-59-0.8427.hdf5\n",
            "5/5 [==============================] - 1s 198ms/step - loss: 0.3343 - accuracy: 0.8427 - val_loss: 1.5926 - val_accuracy: 0.2906\n",
            "Epoch 60/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3214 - accuracy: 0.8542\n",
            "Epoch 60: accuracy improved from 0.84271 to 0.85417, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-60-0.8542.hdf5\n",
            "5/5 [==============================] - 1s 213ms/step - loss: 0.3214 - accuracy: 0.8542 - val_loss: 1.5507 - val_accuracy: 0.3156\n",
            "Epoch 61/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3530 - accuracy: 0.8073\n",
            "Epoch 61: accuracy did not improve from 0.85417\n",
            "5/5 [==============================] - 1s 139ms/step - loss: 0.3530 - accuracy: 0.8073 - val_loss: 1.3845 - val_accuracy: 0.3000\n",
            "Epoch 62/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3181 - accuracy: 0.8573\n",
            "Epoch 62: accuracy improved from 0.85417 to 0.85729, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-62-0.8573.hdf5\n",
            "5/5 [==============================] - 1s 207ms/step - loss: 0.3181 - accuracy: 0.8573 - val_loss: 1.4538 - val_accuracy: 0.3000\n",
            "Epoch 63/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3045 - accuracy: 0.8365\n",
            "Epoch 63: accuracy did not improve from 0.85729\n",
            "5/5 [==============================] - 1s 136ms/step - loss: 0.3045 - accuracy: 0.8365 - val_loss: 1.7603 - val_accuracy: 0.2594\n",
            "Epoch 64/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2769 - accuracy: 0.8698\n",
            "Epoch 64: accuracy improved from 0.85729 to 0.86979, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-64-0.8698.hdf5\n",
            "5/5 [==============================] - 1s 225ms/step - loss: 0.2769 - accuracy: 0.8698 - val_loss: 2.0294 - val_accuracy: 0.2750\n",
            "Epoch 65/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2454 - accuracy: 0.8729\n",
            "Epoch 65: accuracy improved from 0.86979 to 0.87292, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-65-0.8729.hdf5\n",
            "5/5 [==============================] - 1s 201ms/step - loss: 0.2454 - accuracy: 0.8729 - val_loss: 1.9497 - val_accuracy: 0.2969\n",
            "Epoch 66/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2360 - accuracy: 0.8760\n",
            "Epoch 66: accuracy improved from 0.87292 to 0.87604, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-66-0.8760.hdf5\n",
            "5/5 [==============================] - 1s 211ms/step - loss: 0.2360 - accuracy: 0.8760 - val_loss: 2.3227 - val_accuracy: 0.2844\n",
            "Epoch 67/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2386 - accuracy: 0.8865\n",
            "Epoch 67: accuracy improved from 0.87604 to 0.88646, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-67-0.8865.hdf5\n",
            "5/5 [==============================] - 1s 274ms/step - loss: 0.2386 - accuracy: 0.8865 - val_loss: 2.2625 - val_accuracy: 0.3125\n",
            "Epoch 68/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2787 - accuracy: 0.8667\n",
            "Epoch 68: accuracy did not improve from 0.88646\n",
            "5/5 [==============================] - 1s 198ms/step - loss: 0.2787 - accuracy: 0.8667 - val_loss: 2.7945 - val_accuracy: 0.2969\n",
            "Epoch 69/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3364 - accuracy: 0.8344\n",
            "Epoch 69: accuracy did not improve from 0.88646\n",
            "5/5 [==============================] - 1s 259ms/step - loss: 0.3364 - accuracy: 0.8344 - val_loss: 1.8083 - val_accuracy: 0.3344\n",
            "Epoch 70/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2955 - accuracy: 0.8365\n",
            "Epoch 70: accuracy did not improve from 0.88646\n",
            "5/5 [==============================] - 1s 202ms/step - loss: 0.2955 - accuracy: 0.8365 - val_loss: 1.7292 - val_accuracy: 0.3281\n",
            "Epoch 71/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2649 - accuracy: 0.8656\n",
            "Epoch 71: accuracy did not improve from 0.88646\n",
            "5/5 [==============================] - 1s 162ms/step - loss: 0.2649 - accuracy: 0.8656 - val_loss: 1.6363 - val_accuracy: 0.3344\n",
            "Epoch 72/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2358 - accuracy: 0.8875\n",
            "Epoch 72: accuracy improved from 0.88646 to 0.88750, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-72-0.8875.hdf5\n",
            "5/5 [==============================] - 1s 271ms/step - loss: 0.2358 - accuracy: 0.8875 - val_loss: 1.9481 - val_accuracy: 0.3281\n",
            "Epoch 73/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2347 - accuracy: 0.8979\n",
            "Epoch 73: accuracy improved from 0.88750 to 0.89792, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-73-0.8979.hdf5\n",
            "5/5 [==============================] - 1s 248ms/step - loss: 0.2347 - accuracy: 0.8979 - val_loss: 2.0040 - val_accuracy: 0.3063\n",
            "Epoch 74/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2574 - accuracy: 0.8823\n",
            "Epoch 74: accuracy did not improve from 0.89792\n",
            "5/5 [==============================] - 1s 144ms/step - loss: 0.2574 - accuracy: 0.8823 - val_loss: 2.1724 - val_accuracy: 0.3375\n",
            "Epoch 75/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2778 - accuracy: 0.8594\n",
            "Epoch 75: accuracy did not improve from 0.89792\n",
            "5/5 [==============================] - 1s 143ms/step - loss: 0.2778 - accuracy: 0.8594 - val_loss: 1.9113 - val_accuracy: 0.3156\n",
            "Epoch 76/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.3134 - accuracy: 0.8479\n",
            "Epoch 76: accuracy did not improve from 0.89792\n",
            "5/5 [==============================] - 1s 137ms/step - loss: 0.3134 - accuracy: 0.8479 - val_loss: 1.4716 - val_accuracy: 0.3812\n",
            "Epoch 77/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2801 - accuracy: 0.8833\n",
            "Epoch 77: accuracy did not improve from 0.89792\n",
            "5/5 [==============================] - 1s 136ms/step - loss: 0.2801 - accuracy: 0.8833 - val_loss: 1.4815 - val_accuracy: 0.3063\n",
            "Epoch 78/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2461 - accuracy: 0.8885\n",
            "Epoch 78: accuracy did not improve from 0.89792\n",
            "5/5 [==============================] - 1s 141ms/step - loss: 0.2461 - accuracy: 0.8885 - val_loss: 1.4317 - val_accuracy: 0.3531\n",
            "Epoch 79/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2187 - accuracy: 0.9083\n",
            "Epoch 79: accuracy improved from 0.89792 to 0.90833, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-79-0.9083.hdf5\n",
            "5/5 [==============================] - 1s 213ms/step - loss: 0.2187 - accuracy: 0.9083 - val_loss: 1.9267 - val_accuracy: 0.3375\n",
            "Epoch 80/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1942 - accuracy: 0.9177\n",
            "Epoch 80: accuracy improved from 0.90833 to 0.91771, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-80-0.9177.hdf5\n",
            "5/5 [==============================] - 1s 308ms/step - loss: 0.1942 - accuracy: 0.9177 - val_loss: 2.1773 - val_accuracy: 0.3313\n",
            "Epoch 81/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2327 - accuracy: 0.9021\n",
            "Epoch 81: accuracy did not improve from 0.91771\n",
            "5/5 [==============================] - 1s 174ms/step - loss: 0.2327 - accuracy: 0.9021 - val_loss: 2.0848 - val_accuracy: 0.3469\n",
            "Epoch 82/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2357 - accuracy: 0.8969\n",
            "Epoch 82: accuracy did not improve from 0.91771\n",
            "5/5 [==============================] - 1s 167ms/step - loss: 0.2357 - accuracy: 0.8969 - val_loss: 2.1814 - val_accuracy: 0.2656\n",
            "Epoch 83/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2033 - accuracy: 0.9073\n",
            "Epoch 83: accuracy did not improve from 0.91771\n",
            "5/5 [==============================] - 1s 162ms/step - loss: 0.2033 - accuracy: 0.9073 - val_loss: 2.2551 - val_accuracy: 0.3250\n",
            "Epoch 84/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1947 - accuracy: 0.9229\n",
            "Epoch 84: accuracy improved from 0.91771 to 0.92292, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-84-0.9229.hdf5\n",
            "5/5 [==============================] - 1s 228ms/step - loss: 0.1947 - accuracy: 0.9229 - val_loss: 2.4019 - val_accuracy: 0.3438\n",
            "Epoch 85/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1561 - accuracy: 0.9312\n",
            "Epoch 85: accuracy improved from 0.92292 to 0.93125, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-85-0.9312.hdf5\n",
            "5/5 [==============================] - 1s 224ms/step - loss: 0.1561 - accuracy: 0.9312 - val_loss: 2.6517 - val_accuracy: 0.3156\n",
            "Epoch 86/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1740 - accuracy: 0.9250\n",
            "Epoch 86: accuracy did not improve from 0.93125\n",
            "5/5 [==============================] - 1s 156ms/step - loss: 0.1740 - accuracy: 0.9250 - val_loss: 2.5198 - val_accuracy: 0.3219\n",
            "Epoch 87/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2035 - accuracy: 0.9115\n",
            "Epoch 87: accuracy did not improve from 0.93125\n",
            "5/5 [==============================] - 1s 160ms/step - loss: 0.2035 - accuracy: 0.9115 - val_loss: 2.5273 - val_accuracy: 0.3281\n",
            "Epoch 88/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1881 - accuracy: 0.9146\n",
            "Epoch 88: accuracy did not improve from 0.93125\n",
            "5/5 [==============================] - 1s 144ms/step - loss: 0.1881 - accuracy: 0.9146 - val_loss: 3.0708 - val_accuracy: 0.2906\n",
            "Epoch 89/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1732 - accuracy: 0.9187\n",
            "Epoch 89: accuracy did not improve from 0.93125\n",
            "5/5 [==============================] - 1s 161ms/step - loss: 0.1732 - accuracy: 0.9187 - val_loss: 2.5735 - val_accuracy: 0.3406\n",
            "Epoch 90/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1499 - accuracy: 0.9271\n",
            "Epoch 90: accuracy did not improve from 0.93125\n",
            "5/5 [==============================] - 1s 152ms/step - loss: 0.1499 - accuracy: 0.9271 - val_loss: 3.4060 - val_accuracy: 0.2906\n",
            "Epoch 91/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1522 - accuracy: 0.9438\n",
            "Epoch 91: accuracy improved from 0.93125 to 0.94375, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-91-0.9438.hdf5\n",
            "5/5 [==============================] - 1s 230ms/step - loss: 0.1522 - accuracy: 0.9438 - val_loss: 2.8805 - val_accuracy: 0.3313\n",
            "Epoch 92/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1368 - accuracy: 0.9354\n",
            "Epoch 92: accuracy did not improve from 0.94375\n",
            "5/5 [==============================] - 1s 158ms/step - loss: 0.1368 - accuracy: 0.9354 - val_loss: 3.7813 - val_accuracy: 0.2781\n",
            "Epoch 93/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1111 - accuracy: 0.9573\n",
            "Epoch 93: accuracy improved from 0.94375 to 0.95729, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-93-0.9573.hdf5\n",
            "5/5 [==============================] - 1s 264ms/step - loss: 0.1111 - accuracy: 0.9573 - val_loss: 3.8711 - val_accuracy: 0.3031\n",
            "Epoch 94/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1907 - accuracy: 0.9260\n",
            "Epoch 94: accuracy did not improve from 0.95729\n",
            "5/5 [==============================] - 1s 144ms/step - loss: 0.1907 - accuracy: 0.9260 - val_loss: 2.5604 - val_accuracy: 0.3344\n",
            "Epoch 95/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1501 - accuracy: 0.9302\n",
            "Epoch 95: accuracy did not improve from 0.95729\n",
            "5/5 [==============================] - 1s 148ms/step - loss: 0.1501 - accuracy: 0.9302 - val_loss: 2.2662 - val_accuracy: 0.3438\n",
            "Epoch 96/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.2059 - accuracy: 0.9156\n",
            "Epoch 96: accuracy did not improve from 0.95729\n",
            "5/5 [==============================] - 1s 158ms/step - loss: 0.2059 - accuracy: 0.9156 - val_loss: 2.3589 - val_accuracy: 0.3406\n",
            "Epoch 97/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1481 - accuracy: 0.9354\n",
            "Epoch 97: accuracy did not improve from 0.95729\n",
            "5/5 [==============================] - 1s 138ms/step - loss: 0.1481 - accuracy: 0.9354 - val_loss: 2.4533 - val_accuracy: 0.3281\n",
            "Epoch 98/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1206 - accuracy: 0.9479\n",
            "Epoch 98: accuracy did not improve from 0.95729\n",
            "5/5 [==============================] - 1s 152ms/step - loss: 0.1206 - accuracy: 0.9479 - val_loss: 2.9553 - val_accuracy: 0.2969\n",
            "Epoch 99/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1259 - accuracy: 0.9458\n",
            "Epoch 99: accuracy did not improve from 0.95729\n",
            "5/5 [==============================] - 1s 184ms/step - loss: 0.1259 - accuracy: 0.9458 - val_loss: 3.0243 - val_accuracy: 0.3344\n",
            "Epoch 100/100\n",
            "5/5 [==============================] - ETA: 0s - loss: 0.1052 - accuracy: 0.9594\n",
            "Epoch 100: accuracy improved from 0.95729 to 0.95938, saving model to /content/drive/Shareddrives/DEAP EEG/Models Saved/Valence/Raw_MDPI_M1_Heavy/weights-improvement-100-0.9594.hdf5\n",
            "5/5 [==============================] - 1s 240ms/step - loss: 0.1052 - accuracy: 0.9594 - val_loss: 3.6324 - val_accuracy: 0.2937\n"
          ]
        }
      ]
    }
  ]
}