02_explore_gaussian_assumption.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "1086b863-525f-4ccc-ba9d-3c01ed6959d2",
   "metadata": {},
   "source": [
    "<font size=\"+3\"><mark>Limitations of SAX-like methods / The means per segment do not come from a normal distribution</mark></font>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88b04518-40d9-4cf3-97a6-6200fa76f3ae",
   "metadata": {},
   "source": [
    "# Introduction"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "89fd317e-c115-40e1-a729-00d1c964760c",
   "metadata": {},
   "source": [
    "## README"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aa1814e1-2809-4b65-bae3-046c12b36ff8",
   "metadata": {},
   "source": [
    "_Associated GitHub repository: https://github.com/sylvaincom/astride._\n",
    "\n",
    "This notebook investigates the normality assumption on the means per (uniform) segment.\n",
    "\n",
    "It focuses on the 86 data sets that are used in the classification benchmark."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9df79a5-7a63-4518-a697-e43c05885f25",
   "metadata": {},
   "source": [
    "## Configuration parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "484be70b-821b-4714-9117-09d4e9e958a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "IS_COMPUTE = True\n",
    "IS_EXPORT_DF = True\n",
    "IS_SAVE_FIG = True\n",
    "\n",
    "DATE_EXP = \"2023_02_08\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0cc61e5e-ca82-4460-b36a-8160fe4d817e",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a16ae489-2d1c-47a9-aea5-2387919932bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import numpy.ma as ma\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import random\n",
    "from tqdm import tqdm\n",
    "from datetime import timedelta\n",
    "from time import time\n",
    "import plotly.express as px\n",
    "import plotly.figure_factory as ff\n",
    "\n",
    "import ruptures as rpt\n",
    "from scipy.spatial.distance import euclidean\n",
    "from scipy.stats import normaltest, shapiro\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.pipeline import make_pipeline\n",
    "from sklearn.utils import Bunch\n",
    "from tslearn.metrics import dtw\n",
    "from tslearn.preprocessing import TimeSeriesScalerMeanVariance\n",
    "\n",
    "from src.metadata import l_datasets_classif_bench\n",
    "from src.segment_feature import SegmentFeature\n",
    "from src.segmentation import Segmentation\n",
    "from src.symbolic_signal_distance import SymbolicSignalDistance\n",
    "from src.symbolization import Symbolization\n",
    "from src.utils import load_ucr_dataset, create_path\n",
    "from src.utils_gaussian_assumption import run_gaussian_assumption, explore_dataset_means"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "488f3af3-b020-40c1-9b9a-019e5c15fd1c",
   "metadata": {},
   "outputs": [],
   "source": [
    "cwd = Path.cwd()\n",
    "sns.set_theme()\n",
    "plt.rcParams['figure.figsize'] = (10, 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bbeb5d36-7e85-4e41-9e17-834e8e301afe",
   "metadata": {},
   "source": [
    "# Define the data sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "7e9278d3-d52d-48df-97e0-1951c9faa5a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "l_datasets = l_datasets_classif_bench.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2560199d-2105-4067-ac21-67ccbcc5536d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "86\n"
     ]
    }
   ],
   "source": [
    "print(len(l_datasets))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14fc30e7-f2be-43bb-9d0f-bd0445933eb0",
   "metadata": {},
   "source": [
    "# Launch the statistical test for all data sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6f437102-4025-4b27-9243-646d135b4fb4",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 86/86 [03:12<00:00,  2.24s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3min 5s, sys: 4.08 s, total: 3min 9s\n",
      "Wall time: 3min 12s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "if IS_COMPUTE:\n",
    "    d_segment_features_df = dict()\n",
    "    d_normality_test = dict()\n",
    "    print_bool = False\n",
    "\n",
    "    for (i, dataset_name_ucr) in enumerate(tqdm(l_datasets)):\n",
    "        b_get_means_per_segments = run_gaussian_assumption(\n",
    "            dataset_name_ucr=dataset_name_ucr,\n",
    "        )\n",
    "        d_segment_features_df[dataset_name_ucr] = b_get_means_per_segments.segment_features_df\n",
    "        d_normality_test[dataset_name_ucr] = [\n",
    "            b_get_means_per_segments.n_samples,\n",
    "            b_get_means_per_segments.n_segments,\n",
    "            b_get_means_per_segments.p_value,\n",
    "            b_get_means_per_segments.res_H0\n",
    "        ]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92b167de-47e5-4961-8928-63cb042cc49d",
   "metadata": {},
   "source": [
    "# Explore final results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebc3de39-6d2f-4bdb-a6cf-5e63618214f5",
   "metadata": {},
   "source": [
    "## Process results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "14786d1f-bf03-4c5d-bc79-b2815e513568",
   "metadata": {},
   "outputs": [
    {
     "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>dataset</th>\n",
       "      <th>n_samples</th>\n",
       "      <th>n_segments</th>\n",
       "      <th>p_value</th>\n",
       "      <th>normality_assumption</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Adiac</td>\n",
       "      <td>176</td>\n",
       "      <td>32</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>rejected</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ArrowHead</td>\n",
       "      <td>251</td>\n",
       "      <td>32</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>rejected</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Beef</td>\n",
       "      <td>470</td>\n",
       "      <td>64</td>\n",
       "      <td>2.584412e-15</td>\n",
       "      <td>rejected</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BeetleFly</td>\n",
       "      <td>512</td>\n",
       "      <td>64</td>\n",
       "      <td>5.286344e-33</td>\n",
       "      <td>rejected</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BirdChicken</td>\n",
       "      <td>512</td>\n",
       "      <td>64</td>\n",
       "      <td>2.231850e-14</td>\n",
       "      <td>rejected</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       dataset  n_samples  n_segments       p_value normality_assumption\n",
       "0        Adiac        176          32  0.000000e+00             rejected\n",
       "1    ArrowHead        251          32  0.000000e+00             rejected\n",
       "2         Beef        470          64  2.584412e-15             rejected\n",
       "3    BeetleFly        512          64  5.286344e-33             rejected\n",
       "4  BirdChicken        512          64  2.231850e-14             rejected"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if IS_COMPUTE:\n",
    "    df_normality_test = (\n",
    "        pd.DataFrame.from_dict(d_normality_test, orient='index')\n",
    "        .reset_index().rename(columns={\n",
    "            \"index\":\"dataset\",\n",
    "            0:\"n_samples\",\n",
    "            1:\"n_segments\",\n",
    "            2:\"p_value\",\n",
    "            3:\"normality_assumption\"\n",
    "        })\n",
    "    )\n",
    "    display(df_normality_test.head())\n",
    "\n",
    "if IS_EXPORT_DF:\n",
    "    folder = cwd / f\"results/{DATE_EXP}\"\n",
    "    create_path(folder)\n",
    "    df_normality_test.to_csv(folder / \"normality_test.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "252b15ab-bf18-4070-a676-7238e6cb12a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the computed data\n",
    "df_normality_test = pd.read_csv(cwd / f\"results/{DATE_EXP}/normality_test.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "efbac7b0-cfb9-4b4e-8cfe-64733f96d23c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 0 data sets with a NaN p-value.\n",
      "\n",
      "These (eventual) data sets are the following:\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>dataset</th>\n",
       "      <th>n_samples</th>\n",
       "      <th>n_segments</th>\n",
       "      <th>p_value</th>\n",
       "      <th>normality_assumption</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [dataset, n_samples, n_segments, p_value, normality_assumption]\n",
       "Index: []"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Percentage of data sets that respect the normality assumption or not (beware of NaNs):\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "rejected    100.0\n",
       "Name: normality_assumption, dtype: float64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pvalue_na = df_normality_test[\"p_value\"].isna().sum()\n",
    "print(f\"There are {pvalue_na} data sets with a NaN p-value.\")\n",
    "\n",
    "print(\"\\nThese (eventual) data sets are the following:\")\n",
    "display(df_normality_test[df_normality_test[\"p_value\"].isna()])\n",
    "\n",
    "print(\"\\nPercentage of data sets that respect the normality assumption or not (beware of NaNs):\")\n",
    "display(df_normality_test[\"normality_assumption\"].value_counts() / len(df_normality_test) * 100)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a952db09-c268-4c5b-a70b-b907f2c6d15a",
   "metadata": {},
   "source": [
    "<mark>All data sets are rejected!</mark>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f74d383b-559a-4c5f-af21-7a3f232608c4",
   "metadata": {},
   "outputs": [
    {
     "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>dataset</th>\n",
       "      <th>mean_feat</th>\n",
       "      <th>signal_index</th>\n",
       "      <th>segment_start</th>\n",
       "      <th>segment_end</th>\n",
       "      <th>segment_length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>75051</th>\n",
       "      <td>NonInvasiveFetalECGThorax1</td>\n",
       "      <td>-0.636919</td>\n",
       "      <td>1172</td>\n",
       "      <td>503</td>\n",
       "      <td>514</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128724</th>\n",
       "      <td>MixedShapesSmallTrain</td>\n",
       "      <td>-0.936167</td>\n",
       "      <td>2011</td>\n",
       "      <td>319</td>\n",
       "      <td>335</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74652</th>\n",
       "      <td>NonInvasiveFetalECGThorax2</td>\n",
       "      <td>-0.711779</td>\n",
       "      <td>1166</td>\n",
       "      <td>327</td>\n",
       "      <td>339</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135032</th>\n",
       "      <td>StarLightCurves</td>\n",
       "      <td>0.048068</td>\n",
       "      <td>2109</td>\n",
       "      <td>895</td>\n",
       "      <td>911</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36911</th>\n",
       "      <td>InlineSkate</td>\n",
       "      <td>-0.754936</td>\n",
       "      <td>576</td>\n",
       "      <td>1381</td>\n",
       "      <td>1410</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           dataset  mean_feat  signal_index  segment_start  \\\n",
       "75051   NonInvasiveFetalECGThorax1  -0.636919          1172            503   \n",
       "128724       MixedShapesSmallTrain  -0.936167          2011            319   \n",
       "74652   NonInvasiveFetalECGThorax2  -0.711779          1166            327   \n",
       "135032             StarLightCurves   0.048068          2109            895   \n",
       "36911                  InlineSkate  -0.754936           576           1381   \n",
       "\n",
       "        segment_end  segment_length  \n",
       "75051           514              11  \n",
       "128724          335              16  \n",
       "74652           339              12  \n",
       "135032          911              16  \n",
       "36911          1410              29  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if IS_COMPUTE:\n",
    "    \n",
    "    l_segment_features_df = []\n",
    "    for key, value in d_segment_features_df.items():\n",
    "        dataset_name_ucr = key\n",
    "        segment_features_df = value\n",
    "        segment_features_df[\"dataset\"] = dataset_name_ucr\n",
    "        l_segment_features_df.append(segment_features_df)\n",
    "    all_segment_features_df = pd.concat(l_segment_features_df)\n",
    "\n",
    "    l_columns = [\"dataset\"] + [elem for elem in all_segment_features_df.columns.tolist() if elem != \"dataset\"]\n",
    "    all_segment_features_df = all_segment_features_df[l_columns]\n",
    "\n",
    "    display(all_segment_features_df.sample(5))\n",
    "\n",
    "if IS_EXPORT_DF:\n",
    "    folder = cwd / f\"results/{DATE_EXP}\"\n",
    "    create_path(folder)\n",
    "    all_segment_features_df.to_csv(folder / \"all_segment_features_df.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "01c17383-b428-4ff0-b6e2-50b5c81ae9b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the computed data\n",
    "all_segment_features_df = pd.read_csv(cwd / f\"results/{DATE_EXP}/all_segment_features_df.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "20015390-9976-4b29-be59-9d05bdbd14fa",
   "metadata": {},
   "source": [
    "## Explore results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b2f028cc-51dc-4c36-8f82-292a219fc909",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For the Strawberry data set, there are 0.0% of NaN mean values.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>dataset</th>\n",
       "      <th>n_samples</th>\n",
       "      <th>n_segments</th>\n",
       "      <th>p_value</th>\n",
       "      <th>normality_assumption</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>Strawberry</td>\n",
       "      <td>235</td>\n",
       "      <td>32</td>\n",
       "      <td>0.0</td>\n",
       "      <td>rejected</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       dataset  n_samples  n_segments  p_value normality_assumption\n",
       "49  Strawberry        235          32      0.0             rejected"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "explore_dataset_means(\n",
    "    dataset_name_ucr=\"Strawberry\",\n",
    "    all_segment_features_df=all_segment_features_df,\n",
    "    df_normality_test=df_normality_test,\n",
    "    is_save_fig=True,\n",
    "    date_exp=DATE_EXP,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cb123529-43c2-4598-848d-b40d35f472df",
   "metadata": {},
   "source": [
    "*Note*: This corresponds to Figure 2 of the paper."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "02006e3c-2b8a-4968-ae49-094384ee48ca",
   "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": "3.10.6"
  },
  "toc-autonumbering": true,
  "toc-showcode": false,
  "toc-showmarkdowntxt": false
 },
 "nbformat": 4,
 "nbformat_minor": 5
}