chap3-2_multiclass_classification_reuters.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"chap3-2_multiclass_classification_reuters.ipynb","provenance":[],"collapsed_sections":[]},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"markdown","metadata":{"id":"tSGObKqH2b8D","colab_type":"text"},"source":["# Multiclass classification\n","\n","<strong>Abstract</strong>  \n","In this notebook, neural networks will be build for classifying topics of Reuters' newswires. This is a <strong>multiclass classification</strong> example.\n","\n","<strong>Reference</strong>  \n","See pages 78-84 of \"<strong>Deep Learning with Python</strong>\" by Francois Chollet (2018). \n","\n","<strong>Summary (page 84)</strong>  \n","- For a multiclass classification problem, the final activation function should be softmax, which returns a value between 0 and 1.\n","- An appropreate loss function is categorical_crossentropy.\n","- If you want to classify data with N labels, try to avoid using layers whose nodes are less than N. "]},{"cell_type":"code","metadata":{"id":"8VvH2hWskuZk","colab_type":"code","colab":{}},"source":["import numpy as np\n","import matplotlib.pyplot as plt\n","\n","import warnings\n","warnings.filterwarnings(\"ignore\")"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"Gu8maP-lmEag","colab_type":"text"},"source":["## Loading the Reuters dataset"]},{"cell_type":"code","metadata":{"id":"PT9o2b79mHPS","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":85},"outputId":"3f32ac66-d780-4775-f507-d986b216513d","executionInfo":{"status":"ok","timestamp":1577087760364,"user_tz":-60,"elapsed":1545,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["from keras.datasets import reuters\n","\n","(train_data, train_labels), (test_data, test_labels) = reuters.load_data(num_words = 10000)\n","\n","print(\"train_data shape: {}\".format(train_data.shape))\n","print(\"train_labels shape: {}\".format(train_labels.shape))\n","print(\"test_data shape: {}\".format(test_data.shape))\n","print(\"test_labels shape: {}\".format(test_labels.shape))"],"execution_count":31,"outputs":[{"output_type":"stream","text":["train_data shape: (8982,)\n","train_labels shape: (8982,)\n","test_data shape: (2246,)\n","test_labels shape: (2246,)\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"LkW_axRnmQkJ","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":190},"outputId":"4695fe95-e19f-44c5-b948-f5c2a5ad9211","executionInfo":{"status":"ok","timestamp":1577087760365,"user_tz":-60,"elapsed":1231,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["print(\"the first 5 train_data: \\n{}\".format(train_data[:5]))\n","print(\"\\nthe first 5 train_labes: \\n{}\".format(train_labels[:5]))"],"execution_count":32,"outputs":[{"output_type":"stream","text":["the first 5 train_data: \n","[list([1, 2, 2, 8, 43, 10, 447, 5, 25, 207, 270, 5, 3095, 111, 16, 369, 186, 90, 67, 7, 89, 5, 19, 102, 6, 19, 124, 15, 90, 67, 84, 22, 482, 26, 7, 48, 4, 49, 8, 864, 39, 209, 154, 6, 151, 6, 83, 11, 15, 22, 155, 11, 15, 7, 48, 9, 4579, 1005, 504, 6, 258, 6, 272, 11, 15, 22, 134, 44, 11, 15, 16, 8, 197, 1245, 90, 67, 52, 29, 209, 30, 32, 132, 6, 109, 15, 17, 12])\n"," list([1, 3267, 699, 3434, 2295, 56, 2, 7511, 9, 56, 3906, 1073, 81, 5, 1198, 57, 366, 737, 132, 20, 4093, 7, 2, 49, 2295, 2, 1037, 3267, 699, 3434, 8, 7, 10, 241, 16, 855, 129, 231, 783, 5, 4, 587, 2295, 2, 2, 775, 7, 48, 34, 191, 44, 35, 1795, 505, 17, 12])\n"," list([1, 53, 12, 284, 15, 14, 272, 26, 53, 959, 32, 818, 15, 14, 272, 26, 39, 684, 70, 11, 14, 12, 3886, 18, 180, 183, 187, 70, 11, 14, 102, 32, 11, 29, 53, 44, 704, 15, 14, 19, 758, 15, 53, 959, 47, 1013, 15, 14, 19, 132, 15, 39, 965, 32, 11, 14, 147, 72, 11, 180, 183, 187, 44, 11, 14, 102, 19, 11, 123, 186, 90, 67, 960, 4, 78, 13, 68, 467, 511, 110, 59, 89, 90, 67, 1390, 55, 2678, 92, 617, 80, 1274, 46, 905, 220, 13, 4, 346, 48, 235, 629, 5, 211, 5, 1118, 7, 2, 81, 5, 187, 11, 15, 9, 1709, 201, 5, 47, 3615, 18, 478, 4514, 5, 1118, 7, 232, 2, 71, 5, 160, 63, 11, 9, 2, 81, 5, 102, 59, 11, 17, 12])\n"," list([1, 4, 686, 867, 558, 4, 37, 38, 309, 2276, 465, 893, 3541, 114, 2902, 69, 312, 35, 15, 7, 335, 1679, 21, 25, 3675, 2, 3498, 58, 69, 68, 493, 5, 25, 465, 377, 2430, 4, 293, 1172, 739, 4379, 8, 7, 1510, 1131, 13, 899, 6, 4, 990, 309, 415, 4519, 6920, 645, 3916, 791, 5, 4379, 75, 8, 24, 10, 1311, 4677, 5, 344, 756, 7, 2, 231, 9691, 2603, 1413, 43, 509, 43, 68, 327, 5, 2, 3498, 297, 638, 73, 430, 22, 4, 580, 7, 48, 41, 30, 2, 136, 4, 344, 298, 4, 580, 40, 344, 5078, 2, 291, 1488, 10, 3148, 5, 231, 6250, 1308, 5, 8250, 7043, 21, 2, 1622, 990, 309, 415, 265, 5992, 8945, 1149, 9118, 2, 4, 344, 9691, 756, 3729, 2, 4667, 2, 3249, 28, 10, 2190, 24, 77, 41, 682, 10, 4851, 2048, 7, 4, 5540, 2926, 1598, 22, 370, 5954, 7541, 5, 54, 5232, 1685, 2916, 10, 1571, 946, 60, 51, 3249, 5249, 4, 73, 2135, 669, 4, 580, 64, 10, 4280, 6, 2, 25, 482, 35, 150, 377, 2430, 7, 10, 2, 836, 2, 4730, 6920, 5, 4379, 2, 2, 3541, 8, 4, 344, 291, 2, 298, 4228, 6, 2223, 24, 2, 41, 343, 430, 210, 6, 3498, 297, 64, 10, 2281, 455, 5, 7003, 125, 222, 17, 12])\n"," list([1, 8295, 111, 8, 25, 166, 40, 638, 10, 436, 22, 265, 9, 621, 575, 1080, 4742, 1149, 2, 6, 438, 8295, 13, 102, 388, 15, 90, 67, 7, 197, 8295, 8, 4, 270, 416, 23, 527, 6, 2, 4891, 4, 1055, 742, 16, 8, 36, 1480, 6, 2124, 100, 543, 5, 645, 362, 6, 2912, 4, 49, 8, 2, 976, 124, 20, 5, 8295, 80, 9, 100, 362, 543, 395, 61, 44, 20, 8295, 8, 16, 40, 1276, 42, 1436, 166, 415, 6, 888, 4, 116, 9, 40, 3089, 4, 303, 163, 16, 64, 772, 13, 94, 156, 17, 12])]\n","\n","the first 5 train_labes: \n","[3 4 3 4 4]\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"_kjBhkJKmxtM","colab_type":"text"},"source":["Let us perfom data preprocessing:\n","- 0 padding for training and test data so that they have the same length (10000).\n","- one-hot encoding for labels."]},{"cell_type":"code","metadata":{"id":"y-o5J-3jmbsx","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":289},"outputId":"9d0d6785-0b2f-4ce6-b4d5-99d560f0adb6","executionInfo":{"status":"ok","timestamp":1577087761898,"user_tz":-60,"elapsed":1435,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["from keras.utils.np_utils import to_categorical\n","\n","def vectorize_sequences(sequences, dimension=10000):\n","  results = np.zeros((len(sequences),dimension))\n","  for i , sequence in enumerate(sequences):\n","    results[i,sequence] = 1.\n","  \n","  return results\n","\n","x_train = vectorize_sequences(train_data)\n","x_test = vectorize_sequences(test_data)\n","\n","print(\"length of the first training data: \", len(x_train[0]))\n","print(\"length of the second training data: \", len(x_train[1]))\n","print(\"length of the third training data: \", len(x_train[2]))\n","\n","print(\"length of the first test data: \", len(x_test[0]))\n","print(\"length of the second test data: \", len(x_test[1]))\n","print(\"length of the third test data: \", len(x_test[2]))\n","\n","\n","\n","one_hot_train_labels = to_categorical(train_labels)\n","one_hot_test_labels = to_categorical(test_labels)\n","\n","print(\"\\nThe first 3 train_labels:\")\n","print(one_hot_train_labels[:3])\n","\n","print(\"\\nThere are %d classes.\"%len(one_hot_train_labels[0]))"],"execution_count":33,"outputs":[{"output_type":"stream","text":["length of the first training data:  10000\n","length of the second training data:  10000\n","length of the third training data:  10000\n","length of the first test data:  10000\n","length of the second test data:  10000\n","length of the third test data:  10000\n","\n","The first 3 train_labels:\n","[[0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.\n","  0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n"," [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.\n","  0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n"," [0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.\n","  0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]]\n","\n","There are 46 classes.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"jTIY6N1Mpexj","colab_type":"text"},"source":["A validation set will be created from the original training dataset."]},{"cell_type":"code","metadata":{"id":"C54vYxUCpP78","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":85},"outputId":"1fce4828-40a3-43af-9b53-59bf42ebb9dc","executionInfo":{"status":"ok","timestamp":1577087853579,"user_tz":-60,"elapsed":345,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["x_val=x_train[:1000]\n","partial_x_train=x_train[1000:]\n","y_val=one_hot_train_labels[:1000]\n","partial_y_train=one_hot_train_labels[1000:]\n","\n","print(\"x_val shape: \",x_val.shape)\n","print(\"partial_x_train\", partial_x_train.shape)\n","print(\"y_val shape: \",y_val.shape)\n","print(\"partial_y_train\", partial_y_train.shape)"],"execution_count":39,"outputs":[{"output_type":"stream","text":["x_val shape:  (1000, 10000)\n","partial_x_train (7982, 10000)\n","y_val shape:  (1000, 46)\n","partial_y_train (7982, 46)\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"oeMq55qho6vG","colab_type":"text"},"source":["## Neural network 1"]},{"cell_type":"code","metadata":{"id":"OWeivbeMmpBW","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":255},"outputId":"e93ca081-41eb-4840-967c-7400bf0a509f","executionInfo":{"status":"ok","timestamp":1577087861237,"user_tz":-60,"elapsed":421,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["from keras import models\n","from keras import layers\n","\n","model=models.Sequential()\n","model.add(layers.Dense(64,activation='relu',input_shape=(10000,)))\n","model.add(layers.Dense(64,activation='relu'))\n","model.add(layers.Dense(46,activation='softmax'))\n","\n","model.summary()"],"execution_count":40,"outputs":[{"output_type":"stream","text":["Model: \"sequential_5\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","dense_13 (Dense)             (None, 64)                640064    \n","_________________________________________________________________\n","dense_14 (Dense)             (None, 64)                4160      \n","_________________________________________________________________\n","dense_15 (Dense)             (None, 46)                2990      \n","=================================================================\n","Total params: 647,214\n","Trainable params: 647,214\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"w0zfh2pZoqzD","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":904},"outputId":"2c6385bd-a452-4e19-a385-d5e1cfb45644","executionInfo":{"status":"ok","timestamp":1577087885925,"user_tz":-60,"elapsed":23234,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["model.compile(optimizer='rmsprop',\n","                        loss='categorical_crossentropy',\n","                          metrics=['accuracy'])\n","\n","history=model.fit(partial_x_train,\n","                           partial_y_train,\n","                           epochs=20,\n","                           batch_size=512,\n","                           validation_data=(x_val, y_val)) "],"execution_count":41,"outputs":[{"output_type":"stream","text":["Train on 7982 samples, validate on 1000 samples\n","Epoch 1/20\n","WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n","\n","WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n","\n","WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:207: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n","\n","WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:216: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n","\n","WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:223: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n","\n","7982/7982 [==============================] - 2s 299us/step - loss: 2.6662 - acc: 0.4188 - val_loss: 1.8151 - val_acc: 0.6240\n","Epoch 2/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 1.4926 - acc: 0.6857 - val_loss: 1.3316 - val_acc: 0.6910\n","Epoch 3/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 1.0885 - acc: 0.7671 - val_loss: 1.1382 - val_acc: 0.7460\n","Epoch 4/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.8415 - acc: 0.8215 - val_loss: 1.0228 - val_acc: 0.7910\n","Epoch 5/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.6684 - acc: 0.8638 - val_loss: 0.9490 - val_acc: 0.8010\n","Epoch 6/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.5253 - acc: 0.8941 - val_loss: 0.9336 - val_acc: 0.7990\n","Epoch 7/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 0.4277 - acc: 0.9113 - val_loss: 0.8752 - val_acc: 0.8200\n","Epoch 8/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.3445 - acc: 0.9276 - val_loss: 0.8987 - val_acc: 0.8090\n","Epoch 9/20\n","7982/7982 [==============================] - 1s 133us/step - loss: 0.2922 - acc: 0.9384 - val_loss: 0.8847 - val_acc: 0.8150\n","Epoch 10/20\n","7982/7982 [==============================] - 1s 134us/step - loss: 0.2405 - acc: 0.9456 - val_loss: 0.8942 - val_acc: 0.8210\n","Epoch 11/20\n","7982/7982 [==============================] - 1s 132us/step - loss: 0.2058 - acc: 0.9483 - val_loss: 0.9261 - val_acc: 0.8160\n","Epoch 12/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 0.1850 - acc: 0.9520 - val_loss: 0.9087 - val_acc: 0.8180\n","Epoch 13/20\n","7982/7982 [==============================] - 1s 132us/step - loss: 0.1643 - acc: 0.9548 - val_loss: 0.9589 - val_acc: 0.8140\n","Epoch 14/20\n","7982/7982 [==============================] - 1s 132us/step - loss: 0.1518 - acc: 0.9548 - val_loss: 0.9778 - val_acc: 0.8140\n","Epoch 15/20\n","7982/7982 [==============================] - 1s 132us/step - loss: 0.1401 - acc: 0.9555 - val_loss: 0.9698 - val_acc: 0.8250\n","Epoch 16/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.1331 - acc: 0.9562 - val_loss: 0.9765 - val_acc: 0.8190\n","Epoch 17/20\n","7982/7982 [==============================] - 1s 132us/step - loss: 0.1229 - acc: 0.9583 - val_loss: 1.0569 - val_acc: 0.8080\n","Epoch 18/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.1170 - acc: 0.9579 - val_loss: 1.0103 - val_acc: 0.8160\n","Epoch 19/20\n","7982/7982 [==============================] - 1s 133us/step - loss: 0.1157 - acc: 0.9584 - val_loss: 1.0984 - val_acc: 0.7960\n","Epoch 20/20\n","7982/7982 [==============================] - 1s 133us/step - loss: 0.1143 - acc: 0.9579 - val_loss: 1.0657 - val_acc: 0.8030\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"puCC9P0BuzdU","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":279},"outputId":"5c5ca5ab-9ebd-4e1f-e468-2f170dfc0c8e","executionInfo":{"status":"ok","timestamp":1577087927998,"user_tz":-60,"elapsed":1112,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["def visualize_history(history):\n","\n","  history_dict = history.history\n","  loss_values=history_dict['loss']\n","  val_loss_values = history_dict['val_loss']\n","  acc_values= history_dict['acc']\n","  val_acc_values= history_dict['val_acc']\n","\n","  epochs = range(1,len(loss_values)+1)\n","\n","  plt.figure(figsize=(13,4))\n","  plt.subplot(121)\n","  plt.plot(epochs, loss_values, 'bo', label='training loss')\n","  plt.plot(epochs, val_loss_values, 'b', label = 'validation loss')\n","  plt.xlabel('Epochs')\n","  plt.ylabel('Loss')\n","  plt.legend(frameon=False)\n","\n","  plt.subplot(122)\n","  plt.plot(epochs, acc_values, 'bo', label='training acc.')\n","  plt.plot(epochs, val_acc_values, 'b', label = 'validation acc.')\n","  plt.xlabel('Epochs')\n","  plt.ylabel('Accuracy')\n","  plt.legend(frameon=False)\n","\n","  plt.show()\n","\n","visualize_history(history)"],"execution_count":43,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAwgAAAEGCAYAAADBkLJOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAYIkSZKkCQcIkiRJkiYcIEiSJEma+D8RJe2W2W9vqQAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 936x288 with 2 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"wGnjPbfkxWan","colab_type":"text"},"source":["In order to avoid overfitting, a model will be trained for 4 epochs."]},{"cell_type":"code","metadata":{"id":"LW66iX5ww_Mg","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":170},"outputId":"ab85b48a-5db4-4f58-c5bf-51b34fae25eb","executionInfo":{"status":"ok","timestamp":1577088019027,"user_tz":-60,"elapsed":5398,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["model=models.Sequential()\n","model.add(layers.Dense(64,activation='relu',input_shape=(10000,)))\n","model.add(layers.Dense(64,activation='relu'))\n","model.add(layers.Dense(46,activation='softmax'))\n","\n","model.compile(optimizer='rmsprop',\n","                        loss='categorical_crossentropy',\n","                          metrics=['accuracy'])\n","\n","history=model.fit(partial_x_train,\n","                           partial_y_train,\n","                           epochs=4,\n","                           batch_size=512,\n","                           validation_data=(x_val, y_val)) "],"execution_count":44,"outputs":[{"output_type":"stream","text":["Train on 7982 samples, validate on 1000 samples\n","Epoch 1/4\n","7982/7982 [==============================] - 1s 187us/step - loss: 2.5975 - acc: 0.4813 - val_loss: 1.7412 - val_acc: 0.6150\n","Epoch 2/4\n","7982/7982 [==============================] - 1s 131us/step - loss: 1.4385 - acc: 0.6931 - val_loss: 1.3443 - val_acc: 0.7000\n","Epoch 3/4\n","7982/7982 [==============================] - 1s 131us/step - loss: 1.0767 - acc: 0.7699 - val_loss: 1.1546 - val_acc: 0.7480\n","Epoch 4/4\n","7982/7982 [==============================] - 1s 131us/step - loss: 0.8499 - acc: 0.8213 - val_loss: 1.0684 - val_acc: 0.7750\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"jq_-YgddxVTp","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":51},"outputId":"39f3b948-d3bb-4c4d-ec07-f80dd880cd60","executionInfo":{"status":"ok","timestamp":1577088098337,"user_tz":-60,"elapsed":926,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["results = model.evaluate(x_test, one_hot_test_labels)\n","results"],"execution_count":45,"outputs":[{"output_type":"stream","text":["2246/2246 [==============================] - 0s 86us/step\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[1.1314735325552583, 0.7502226180140715]"]},"metadata":{"tags":[]},"execution_count":45}]},{"cell_type":"markdown","metadata":{"id":"LIFixRCFyGYt","colab_type":"text"},"source":["It is possible to get a probability distribution."]},{"cell_type":"code","metadata":{"id":"iZuVlKLXxpjG","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":517},"outputId":"58a3bf71-8163-4dea-9034-57a769d775a5","executionInfo":{"status":"ok","timestamp":1577089024100,"user_tz":-60,"elapsed":851,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["predictions = model.predict(x_test)\n","\n","print(\"predictions shape: \", predictions.shape)\n","print(\"the first prediction: \", predictions[0])\n","print(\"sum  = %0.3f (expected to be 1)\"%np.sum(predictions[0]))\n","print(\"maximum at index\", np.argmax(predictions[0]))\n","print(\"the first element in test_labels:\", test_labels[0])\n","\n","plt.figure(figsize=(11,4))\n","plt.plot(predictions[0],'k')\n","plt.xlabel('label')\n","plt.ylabel('probability')\n","plt.show()"],"execution_count":62,"outputs":[{"output_type":"stream","text":["predictions shape:  (2246, 46)\n","the first prediction:  [3.0325892e-04 4.5530135e-03 3.4539826e-04 7.0059621e-01 2.1744494e-01\n"," 3.1774596e-04 3.1271685e-04 6.1887095e-04 1.5340243e-02 1.8954208e-03\n"," 3.8776370e-03 4.5942036e-03 5.6942611e-04 6.3122786e-03 4.2012444e-04\n"," 7.7060504e-05 9.4556995e-03 1.8336707e-03 5.1925466e-03 6.0380050e-03\n"," 8.1320042e-03 3.1287093e-03 5.0499406e-05 5.2658736e-04 1.0214527e-03\n"," 1.6122588e-03 4.0149505e-04 5.2320123e-05 4.8988219e-04 2.7684524e-04\n"," 2.8361971e-04 1.5378413e-04 1.6635944e-04 7.5551070e-05 4.5213895e-04\n"," 5.0527589e-05 9.9081872e-04 1.1403585e-04 4.2570272e-04 3.8500485e-04\n"," 4.4946489e-04 2.6584594e-04 9.2118855e-05 2.1172703e-04 3.5143694e-05\n"," 5.7671034e-05]\n","sum  = 1.000 (expected to be 1)\n","maximum at index 3\n","the first element in test_labels: 3\n"],"name":"stdout"},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAApkAAAEGCAYAAAAud45ZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n","text/plain":["<Figure size 792x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"UAQsV3iK1kZ1","colab_type":"text"},"source":["## The importance of having sufficiently large intermediate layers"]},{"cell_type":"markdown","metadata":{"id":"yOAM5whR1tLU","colab_type":"text"},"source":["Since the output laer is 46 dimensions, you should avoid intermediate layers whose nodes are less than 46. In the next cell, the number of nodes is intentionally set to be less than 46."]},{"cell_type":"code","metadata":{"id":"g1bRTWWpx3QY","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":714},"outputId":"22038813-1937-4581-e13c-6bec3e7649b4","executionInfo":{"status":"ok","timestamp":1577089154700,"user_tz":-60,"elapsed":23194,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["model=models.Sequential()\n","model.add(layers.Dense(64,activation='relu',input_shape=(10000,)))\n","model.add(layers.Dense(4,activation='relu'))\n","model.add(layers.Dense(46,activation='softmax'))\n","\n","model.compile(optimizer='rmsprop',\n","                        loss='categorical_crossentropy',\n","                          metrics=['accuracy'])\n","\n","history=model.fit(partial_x_train,\n","                           partial_y_train,\n","                           epochs=20,\n","                           batch_size=512,\n","                           validation_data=(x_val, y_val)) "],"execution_count":63,"outputs":[{"output_type":"stream","text":["Train on 7982 samples, validate on 1000 samples\n","Epoch 1/20\n","7982/7982 [==============================] - 3s 329us/step - loss: 3.4815 - acc: 0.1175 - val_loss: 3.1473 - val_acc: 0.3790\n","Epoch 2/20\n","7982/7982 [==============================] - 1s 133us/step - loss: 2.8904 - acc: 0.4078 - val_loss: 2.6699 - val_acc: 0.4090\n","Epoch 3/20\n","7982/7982 [==============================] - 1s 134us/step - loss: 2.4296 - acc: 0.4162 - val_loss: 2.3154 - val_acc: 0.4130\n","Epoch 4/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 2.0929 - acc: 0.4206 - val_loss: 2.0692 - val_acc: 0.4160\n","Epoch 5/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 1.8530 - acc: 0.5219 - val_loss: 1.8860 - val_acc: 0.5840\n","Epoch 6/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 1.6622 - acc: 0.6347 - val_loss: 1.7660 - val_acc: 0.6020\n","Epoch 7/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 1.5089 - acc: 0.6584 - val_loss: 1.6436 - val_acc: 0.6360\n","Epoch 8/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 1.3923 - acc: 0.6724 - val_loss: 1.5926 - val_acc: 0.6310\n","Epoch 9/20\n","7982/7982 [==============================] - 1s 129us/step - loss: 1.3004 - acc: 0.6803 - val_loss: 1.5247 - val_acc: 0.6460\n","Epoch 10/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 1.2235 - acc: 0.6859 - val_loss: 1.4877 - val_acc: 0.6480\n","Epoch 11/20\n","7982/7982 [==============================] - 1s 129us/step - loss: 1.1586 - acc: 0.6888 - val_loss: 1.4859 - val_acc: 0.6460\n","Epoch 12/20\n","7982/7982 [==============================] - 1s 128us/step - loss: 1.1008 - acc: 0.6914 - val_loss: 1.4415 - val_acc: 0.6540\n","Epoch 13/20\n","7982/7982 [==============================] - 1s 132us/step - loss: 1.0431 - acc: 0.6968 - val_loss: 1.4237 - val_acc: 0.6520\n","Epoch 14/20\n","7982/7982 [==============================] - 1s 133us/step - loss: 0.9922 - acc: 0.7067 - val_loss: 1.4032 - val_acc: 0.6580\n","Epoch 15/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.9398 - acc: 0.7167 - val_loss: 1.4346 - val_acc: 0.6520\n","Epoch 16/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.8945 - acc: 0.7425 - val_loss: 1.3968 - val_acc: 0.6720\n","Epoch 17/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 0.8538 - acc: 0.7712 - val_loss: 1.4029 - val_acc: 0.6940\n","Epoch 18/20\n","7982/7982 [==============================] - 1s 130us/step - loss: 0.8158 - acc: 0.7895 - val_loss: 1.4020 - val_acc: 0.6980\n","Epoch 19/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 0.7784 - acc: 0.8062 - val_loss: 1.4212 - val_acc: 0.7040\n","Epoch 20/20\n","7982/7982 [==============================] - 1s 131us/step - loss: 0.7438 - acc: 0.8148 - val_loss: 1.4332 - val_acc: 0.7080\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"Oy3VnfQO1mOl","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":279},"outputId":"888acf90-889c-42a7-90f7-50ba895e83ec","executionInfo":{"status":"ok","timestamp":1577089287196,"user_tz":-60,"elapsed":1025,"user":{"displayName":"Kotaro Sonoda","photoUrl":"","userId":"14975548384331586092"}}},"source":["visualize_history(history)"],"execution_count":64,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAwgAAAEGCAYAAADBkLJOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCJIkSZJ6HCBIkiRJ6nGAIEmSJKnnPwygSf4k7CisAAAAAElFTkSuQmCC\n","text/plain":["<Figure size 936x288 with 2 Axes>"]},"metadata":{"tags":[]}}]}]}