tutorials/waveform_unsupervised-FRUFS.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "48e8bcc7",
   "metadata": {},
   "source": [
    "## Get all the imports here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c565b7e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import random\n",
    "from sklearn.metrics.cluster import normalized_mutual_info_score, adjusted_rand_score\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.cluster import KMeans, AgglomerativeClustering\n",
    "from tqdm.notebook import tqdm\n",
    "import joblib\n",
    "from joblib import Parallel, delayed\n",
    "import contextlib\n",
    "np.random.seed(27)\n",
    "random.seed(27)\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from lightgbm import LGBMRegressor\n",
    "from collections import Counter\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from FRUFS import FRUFS\n",
    "from matplotlib.pyplot import figure"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cce21b89",
   "metadata": {},
   "source": [
    "## Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7a78368f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Data Shape: (4000, 21)\n",
      "Test Data Shape: (1000, 21)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.23</td>\n",
       "      <td>-1.56</td>\n",
       "      <td>-1.75</td>\n",
       "      <td>-0.28</td>\n",
       "      <td>0.60</td>\n",
       "      <td>2.22</td>\n",
       "      <td>0.85</td>\n",
       "      <td>0.21</td>\n",
       "      <td>-0.20</td>\n",
       "      <td>0.89</td>\n",
       "      <td>...</td>\n",
       "      <td>2.89</td>\n",
       "      <td>7.75</td>\n",
       "      <td>4.59</td>\n",
       "      <td>3.15</td>\n",
       "      <td>5.12</td>\n",
       "      <td>3.32</td>\n",
       "      <td>1.20</td>\n",
       "      <td>0.24</td>\n",
       "      <td>-0.56</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.69</td>\n",
       "      <td>2.43</td>\n",
       "      <td>0.61</td>\n",
       "      <td>2.08</td>\n",
       "      <td>2.30</td>\n",
       "      <td>3.25</td>\n",
       "      <td>5.52</td>\n",
       "      <td>4.55</td>\n",
       "      <td>2.97</td>\n",
       "      <td>2.22</td>\n",
       "      <td>...</td>\n",
       "      <td>1.24</td>\n",
       "      <td>1.89</td>\n",
       "      <td>1.88</td>\n",
       "      <td>-1.34</td>\n",
       "      <td>0.83</td>\n",
       "      <td>1.41</td>\n",
       "      <td>1.78</td>\n",
       "      <td>0.60</td>\n",
       "      <td>2.42</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.12</td>\n",
       "      <td>-0.94</td>\n",
       "      <td>1.29</td>\n",
       "      <td>2.59</td>\n",
       "      <td>2.42</td>\n",
       "      <td>3.55</td>\n",
       "      <td>4.94</td>\n",
       "      <td>3.25</td>\n",
       "      <td>1.90</td>\n",
       "      <td>2.07</td>\n",
       "      <td>...</td>\n",
       "      <td>2.50</td>\n",
       "      <td>0.12</td>\n",
       "      <td>1.41</td>\n",
       "      <td>2.78</td>\n",
       "      <td>0.64</td>\n",
       "      <td>0.62</td>\n",
       "      <td>-0.01</td>\n",
       "      <td>-0.79</td>\n",
       "      <td>-0.12</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.86</td>\n",
       "      <td>0.29</td>\n",
       "      <td>2.19</td>\n",
       "      <td>-0.02</td>\n",
       "      <td>1.13</td>\n",
       "      <td>2.51</td>\n",
       "      <td>2.37</td>\n",
       "      <td>5.45</td>\n",
       "      <td>5.45</td>\n",
       "      <td>4.84</td>\n",
       "      <td>...</td>\n",
       "      <td>2.58</td>\n",
       "      <td>1.40</td>\n",
       "      <td>1.24</td>\n",
       "      <td>1.41</td>\n",
       "      <td>1.07</td>\n",
       "      <td>-1.43</td>\n",
       "      <td>2.84</td>\n",
       "      <td>-1.18</td>\n",
       "      <td>1.12</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.16</td>\n",
       "      <td>0.37</td>\n",
       "      <td>0.40</td>\n",
       "      <td>-0.59</td>\n",
       "      <td>2.66</td>\n",
       "      <td>1.00</td>\n",
       "      <td>2.69</td>\n",
       "      <td>4.06</td>\n",
       "      <td>5.34</td>\n",
       "      <td>3.53</td>\n",
       "      <td>...</td>\n",
       "      <td>4.30</td>\n",
       "      <td>1.84</td>\n",
       "      <td>1.73</td>\n",
       "      <td>0.21</td>\n",
       "      <td>-0.18</td>\n",
       "      <td>0.13</td>\n",
       "      <td>-0.21</td>\n",
       "      <td>-0.80</td>\n",
       "      <td>-0.68</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     0     1     2     3     4     5     6     7     8     9   ...    12  \\\n",
       "0 -1.23 -1.56 -1.75 -0.28  0.60  2.22  0.85  0.21 -0.20  0.89  ...  2.89   \n",
       "1 -0.69  2.43  0.61  2.08  2.30  3.25  5.52  4.55  2.97  2.22  ...  1.24   \n",
       "2 -0.12 -0.94  1.29  2.59  2.42  3.55  4.94  3.25  1.90  2.07  ...  2.50   \n",
       "3  0.86  0.29  2.19 -0.02  1.13  2.51  2.37  5.45  5.45  4.84  ...  2.58   \n",
       "4  1.16  0.37  0.40 -0.59  2.66  1.00  2.69  4.06  5.34  3.53  ...  4.30   \n",
       "\n",
       "     13    14    15    16    17    18    19    20  21  \n",
       "0  7.75  4.59  3.15  5.12  3.32  1.20  0.24 -0.56   2  \n",
       "1  1.89  1.88 -1.34  0.83  1.41  1.78  0.60  2.42   1  \n",
       "2  0.12  1.41  2.78  0.64  0.62 -0.01 -0.79 -0.12   0  \n",
       "3  1.40  1.24  1.41  1.07 -1.43  2.84 -1.18  1.12   1  \n",
       "4  1.84  1.73  0.21 -0.18  0.13 -0.21 -0.80 -0.68   1  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"../data/waveform/waveform.data\", header=None)\n",
    "X, Y = df.drop([21], axis=1).values, df[21].values\n",
    "num_classes = len(set(Y))\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2)\n",
    "\n",
    "\n",
    "# Check out the data shapes\n",
    "print(\"Train Data Shape:\", X_train.shape)\n",
    "print(\"Test Data Shape:\", X_test.shape)\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2ceb6185",
   "metadata": {},
   "source": [
    "## Grab a baseline score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e7ade8de",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Baseline NMI Score: 0.3820361000671423\n"
     ]
    }
   ],
   "source": [
    "# We are going to use the K-Means algorithm\n",
    "model_kmeans = AgglomerativeClustering(n_clusters=num_classes)\n",
    "preds = model_kmeans.fit_predict(X_train)\n",
    "\n",
    "# We are going to use the NMI metric to measure the quality/performance of the clustering \n",
    "baseline_score = normalized_mutual_info_score(Y_train, preds)\n",
    "print(\"Baseline NMI Score:\", baseline_score)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8d9c4dd",
   "metadata": {},
   "source": [
    "## Run the unsupervised feature selection method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "560ddaf1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Progress bar: 100%|████████| 21/21 [00:06<00:00,  3.37it/s]\n"
     ]
    }
   ],
   "source": [
    "model = FRUFS(model_r=LGBMRegressor(), k=0.72, n_jobs=-1, verbose=0, random_state=27)\n",
    "X_train_pruned = model.fit_transform(X_train)\n",
    "X_test_pruned = model.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26427132",
   "metadata": {},
   "source": [
    "## Check out the performance of the selected features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "66cfaec8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NMI Score with feature selection 0.39673691763610214\n"
     ]
    }
   ],
   "source": [
    "model_kmeans = AgglomerativeClustering(n_clusters=num_classes)\n",
    "preds = model_kmeans.fit_predict(X_train_pruned)\n",
    "print(\"NMI Score with feature selection\", normalized_mutual_info_score(Y_train, preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b2664f54",
   "metadata": {},
   "source": [
    "## Plot out the feature importances. The blue dashed line indicates the cut-off point you have set.\n",
    "\n",
    "## Note: You can set K in FRUFS by inspecting the feature importance "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "f5e71cce",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x1000 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figure(figsize=(8, 10), dpi=100)\n",
    "model.feature_importance()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4e52578a",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}