benchmarking_result.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.53125\n"
     ]
    }
   ],
   "source": [
    "import dill as pickle\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from test_utils import extract_labels, get_optimizer, create_dirs\n",
    "import os \n",
    "\n",
    "import random\n",
    "import numpy as np\n",
    "\n",
    "import sys\n",
    "sys.path.append('/Users/matsumoton/Git/digen')\n",
    "from digen import Benchmark"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from os import makedirs\n",
    "from matplotlib.lines import Line2D\n",
    "\n",
    "\n",
    "def get_files_name(directory_name, inputname, i, evpop, evgen):\n",
    "    cx_name = f\"{directory_name}{inputname}_run_{i}_pop{evpop}_gen{evgen}_crossover.csv\"\n",
    "    mut_name = f\"{directory_name}{inputname}_run_{i}_pop{evpop}_gen{evgen}_mutation.csv\"\n",
    "    return cx_name, mut_name\n",
    "\n",
    "\n",
    "def determine_better(row):\n",
    "    if row.offspring <= row.parent1 and row.offspring <= row.parent2:\n",
    "        return 'worse'\n",
    "    if row.offspring <= row.parent1 or row.offspring <= row.parent2:\n",
    "        return 'better than one'\n",
    "    return 'better than two'\n",
    "\n",
    "\n",
    "def count_improvement(cx_dfs, mut_dfs):\n",
    "    cx_counters = [{'worse': 0, 'better than one': 0, 'better than two': 0} for _ in range(len(cx_dfs))]\n",
    "    for cx_c, cx_df in zip(cx_counters, cx_dfs):\n",
    "        for row in cx_df.itertuples(index=False):\n",
    "            offspring_better = determine_better(row)\n",
    "            cx_c[offspring_better] += 1\n",
    "\n",
    "    mut_counters = [{'worse': 0, 'better': 0} for _ in range(len(mut_dfs))]\n",
    "    for mut_c, mut_df in zip(mut_counters, mut_dfs):\n",
    "        for row in mut_df.itertuples(index=False):\n",
    "            if row.offspring <= row.parent:\n",
    "                mut_c['worse'] += 1\n",
    "            else:\n",
    "                mut_c['better'] += 1\n",
    "\n",
    "    return cx_counters, mut_counters\n",
    "\n",
    "\n",
    "def compute_p_each_generation(cx_dfs, mut_dfs, max_gen):\n",
    "    cx_ps = [[] for _ in range(1, max_gen)]\n",
    "    mut_ps = [[] for _ in range(1, max_gen)]\n",
    "    for g in range(1, max_gen):\n",
    "        for cx_df in cx_dfs:\n",
    "            tmp_df = cx_df.loc[cx_df['gen'] == g]\n",
    "            if tmp_df.shape[0] != 0:\n",
    "                n_increment = 0\n",
    "                for row in tmp_df.itertuples(index=False):\n",
    "                    offspring_better = determine_better(row)\n",
    "                    if offspring_better != 'worse':\n",
    "                        n_increment += 1\n",
    "                p = n_increment / tmp_df.shape[0]\n",
    "                cx_ps[g - 1].append(p * 100)\n",
    "        for mut_df in mut_dfs:\n",
    "            tmp_df = mut_df.loc[mut_df['gen'] == g]\n",
    "            if tmp_df.shape[0] != 0:\n",
    "                n_increment = 0\n",
    "                for row in tmp_df.itertuples(index=False):\n",
    "                    n_increment += 1 if row.offspring >= row.parent else 0\n",
    "                p = n_increment / tmp_df.shape[0]\n",
    "                mut_ps[g - 1].append(p * 100)\n",
    "    return cx_ps, mut_ps\n",
    "\n",
    "\n",
    "def compute_aggregate_metrics(counters):\n",
    "    keys = [k for k in counters[0]]\n",
    "    aggregate = {k: [] for k in keys}\n",
    "    for cx_c in counters:\n",
    "        for k in keys:\n",
    "            aggregate[k].append(cx_c[k])\n",
    "    aggregate_metrics = {k: {'mean': np.mean(aggregate[k]),\n",
    "                             'std': np.std(aggregate[k]),\n",
    "                             'median': np.median(aggregate[k])}\n",
    "                         for k in aggregate}\n",
    "    return aggregate_metrics\n",
    "\n",
    "\n",
    "def aggregate_metrics(cx_counters, mut_counters):\n",
    "    cx_metrics = compute_aggregate_metrics(cx_counters)\n",
    "    mut_metrics = compute_aggregate_metrics(mut_counters)\n",
    "    return cx_metrics, mut_metrics\n",
    "\n",
    "\n",
    "def plot_improvements(c_cx, c_mut, title, save_directory, filename):\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    plt.bar([1, 2, 3],\n",
    "            [c_cx['worse']['mean'], c_cx['better than one']['mean'], c_cx['better than two']['mean']],\n",
    "            alpha=0.9)\n",
    "    # plot medians\n",
    "    for x, v in zip([1, 2, 3],\n",
    "                    [c_cx['worse']['median'], c_cx['better than one']['median'], c_cx['better than two']['median']]):\n",
    "        plt.plot([x - 0.35, x + 0.35], [v, v], lw=2.5, color='orange', alpha=0.9, zorder=11)\n",
    "    # plot stds\n",
    "    for x, m, std in zip([1, 2, 3],\n",
    "                         [c_cx['worse']['mean'], c_cx['better than one']['mean'], c_cx['better than two']['mean']],\n",
    "                         [c_cx['worse']['std'], c_cx['better than one']['std'], c_cx['better than two']['std']]):\n",
    "        plt.plot([x - 0.1, x + 0.1], [m - std, m - std], lw=1.5, ls='-', color='black', alpha=0.8, zorder=10)\n",
    "        plt.plot([x - 0.1, x + 0.1], [m + std, m + std], lw=1.5, ls='-', color='black', alpha=0.8, zorder=10)\n",
    "        plt.plot([x, x], [m - std, m + std], ls='-', lw=1.5, color='black', alpha=0.8, zorder=10)\n",
    "    plt.xticks([1, 2, 3],\n",
    "               [\"No improvement\", \"Improved one parent\", \"Improved both parents\"],\n",
    "               rotation=30, fontsize=18)\n",
    "    plt.yticks(fontsize=18)\n",
    "    plt.title(f\"{title} - Crossover\", fontsize=26)\n",
    "    plt.legend(handles=[Line2D([0], [0], lw=2.5, label='Mean'),\n",
    "                        Line2D([0], [0], color='orange', lw=2.5, label='Median'),\n",
    "                        Line2D([0], [0], color='black', lw=1.5, label='Standard Deviation')],\n",
    "               fontsize=14)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(f\"{save_directory}/images/pdf/{filename}_cx.pdf\")\n",
    "    plt.savefig(f\"{save_directory}/images/png/{filename}_cx.png\")\n",
    "\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    plt.bar([1, 2],\n",
    "            [c_mut['worse']['mean'], c_mut['better']['mean']],\n",
    "            alpha=0.9)\n",
    "    # plot medians\n",
    "    for x, v in zip([1, 2], [c_mut['worse']['median'], c_mut['better']['median']]):\n",
    "        plt.plot([x - 0.35, x + 0.35], [v, v], lw=2.5, color='orange', alpha=0.9, zorder=11)\n",
    "    # plot stds\n",
    "    for x, m, std in zip([1, 2],\n",
    "                         [c_mut['worse']['mean'], c_mut['better']['mean']],\n",
    "                         [c_mut['worse']['std'], c_mut['better']['std']]):\n",
    "        plt.plot([x - 0.1, x + 0.1], [m - std, m - std], lw=1.5, ls='-', color='black', alpha=0.8, zorder=10)\n",
    "        plt.plot([x - 0.1, x + 0.1], [m + std, m + std], lw=1.5, ls='-', color='black', alpha=0.8, zorder=10)\n",
    "        plt.plot([x, x], [m - std, m + std], ls='-', lw=1.5, color='black', alpha=0.8, zorder=10)\n",
    "    plt.xticks([1, 2], [\"No improvement\", \"Improved\"],\n",
    "               rotation=30, fontsize=18)\n",
    "    plt.yticks(fontsize=18)\n",
    "    plt.title(f\"{title} - Mutation\", fontsize=26)\n",
    "    plt.legend(handles=[Line2D([0], [0], lw=2.5, label='Mean'),\n",
    "                        Line2D([0], [0], color='orange', lw=2.5, label='Median'),\n",
    "                        Line2D([0], [0], color='black', lw=1.5, label='Standard Deviation')],\n",
    "               fontsize=18)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(f\"{save_directory}/images/pdf/{filename}_mut.pdf\")\n",
    "    plt.savefig(f\"{save_directory}/images/png/{filename}_mut.png\")\n",
    "\n",
    "\n",
    "def percentage_metrics(ps):\n",
    "    means = []\n",
    "    stds = []\n",
    "    for gen_p in ps:\n",
    "        means.append(np.mean(gen_p))\n",
    "        stds.append(np.std(gen_p))\n",
    "    return means, stds\n",
    "\n",
    "\n",
    "def plot_percentage_improvement_vs_generation(cx_p_improvement, mut_p_improvement, max_gen,\n",
    "                                              title, save_directory, filename):\n",
    "    cx_mean, cx_std = percentage_metrics(cx_p_improvement)\n",
    "    mut_mean, mut_std = percentage_metrics(mut_p_improvement)\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    plt.plot([i for i in range(1, max_gen)], cx_mean, color='dimgrey', lw=2.5, label=\"Crossover\")\n",
    "    plt.plot([i for i in range(1, max_gen)], mut_mean, color='darkgreen', lw=2.5, label=\"Mutation\")\n",
    "    plt.xticks(fontsize=14)\n",
    "    plt.yticks(fontsize=18)\n",
    "    plt.xlabel(\"Generation\", fontsize=20)\n",
    "    plt.ylabel(\"Improvement percentage\", fontsize=20)\n",
    "    plt.xlim(1, max_gen - 1)\n",
    "    plt.ylim(0)\n",
    "    plt.title(f\"{title}\", fontsize=30)\n",
    "    plt.legend(fontsize=18)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(f\"{save_directory}/images/pdf/{filename}_percentage.pdf\")\n",
    "    plt.savefig(f\"{save_directory}/images/png/{filename}_percentage.png\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from os.path import exists\n",
    "\n",
    "directoryevs = [\"baseline\",\"gi_crossover\"]\n",
    "\n",
    "for j in [2,4,7,14,23,24,25,27,28,30,32,35,40]:\n",
    "    numberofruns = 10\n",
    "    cx_dfs = []\n",
    "    mut_dfs = []\n",
    "    directory = \"/Users/matsumoton/Git/results_pop40_gen15_gi_crossover/offspring_generation_test/\"\n",
    "    inputname = \"digen\" + str(j)\n",
    "    file_exists = exists(f\"{directory}{inputname}_run_0_pop40_gen40_crossover.csv\")\n",
    "    if not file_exists:\n",
    "        continue\n",
    "    for i in range(numberofruns):\n",
    "        cx_df_name, mut_df_name = get_files_name(directory,\"digen\" + str(j), i, 40, 40)\n",
    "        cx_dfs.append(pd.read_csv(cx_df_name, sep=','))\n",
    "        mut_dfs.append(pd.read_csv(mut_df_name, sep=','))\n",
    "    try:\n",
    "        makedirs(f\"{directory}images/pdf/\")\n",
    "        makedirs(f\"{directory}images/png/\")\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "    cx_counters, mut_counters = count_improvement(cx_dfs, mut_dfs)\n",
    "    cx_metrics, mut_metrics = aggregate_metrics(cx_counters, mut_counters)\n",
    "    print(cx_metrics)\n",
    "    print(mut_metrics)\n",
    "    plot_improvements(cx_metrics, mut_metrics, \"digen\"+str(j), directory, \"digen\"+str(j))\n",
    "\n",
    "    cx_p_improvement, mut_p_improvement = compute_p_each_generation(cx_dfs, mut_dfs, 15)\n",
    "    plot_percentage_improvement_vs_generation(cx_p_improvement, mut_p_improvement, 15,\n",
    "                                                \"digen\"+str(j), directory, \"digen\"+str(j))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "import argparse\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from os import makedirs\n",
    "from scipy.stats import wilcoxon\n",
    "from statannot import add_stat_annotation\n",
    "\n",
    "\n",
    "def get_file_name(directory_name, inputname, i, evpop, evgen):\n",
    "    df_name = f\"{directory_name}{inputname}_run_{i}_evolution_pop{evpop}_gen{evgen}.csv\"\n",
    "    return df_name\n",
    "\n",
    "\n",
    "def extract_bests(ev_dfs, nev_dfs):\n",
    "    # extract bests of each generation of evolutions\n",
    "    max_gen = max(ev_dfs[0]['generation'])\n",
    "    bests_ev = [[] for _ in range(max_gen + 1)]\n",
    "    for df in ev_dfs:\n",
    "        for gen in range(max_gen + 1):\n",
    "            gen_scores = df.loc[df['generation'] == gen]['score']\n",
    "            bests_ev[gen].append(max(gen_scores))\n",
    "    bests_ev = np.array(bests_ev)\n",
    "    # extract bests of each random sampling\n",
    "    bests_nev = []\n",
    "    for df in nev_dfs:\n",
    "        bests_nev.append(max(df['score']))\n",
    "    bests_nev = np.array(bests_nev)\n",
    "    return bests_ev, bests_nev\n",
    "\n",
    "\n",
    "def compute_medians_bests(bests_ev, bests_nev):\n",
    "    # median for evolution\n",
    "    medians_ev = np.median(bests_ev, axis=1)\n",
    "    # median for no evolution dfs\n",
    "    median_nev = np.median(bests_nev)\n",
    "    return medians_ev, median_nev\n",
    "\n",
    "\n",
    "def plot_boxplots(last_gen_bests, random_bests, score, configname, title, save_directory, filename):\n",
    "    df_ev_last_gen = pd.DataFrame({'score': last_gen_bests})\n",
    "    df_ev_last_gen['type'] = configname\n",
    "    df_nev = pd.DataFrame({'score': random_bests})\n",
    "    df_nev['type'] = 'Random sampling'\n",
    "    df = pd.concat((df_ev_last_gen, df_nev))\n",
    "\n",
    "    _, pvalue = wilcoxon(last_gen_bests, random_bests,\n",
    "                         zero_method='wilcox', correction=False,\n",
    "                         alternative='two-sided', mode='auto')\n",
    "    print(f\"Wilcoxon Two-sided pvalue: {pvalue}\")\n",
    "\n",
    "    fig, ax = plt.subplots(1, 1, figsize=(8, 8))\n",
    "    sns.despine(offset=10, trim=False)\n",
    "    sns.boxplot(x='type', y='score', data=df, ax=ax, showmeans=False, dodge=False)\n",
    "    add_stat_annotation(ax, data=df, x='type', y='score', perform_stat_test=False,\n",
    "                        box_pairs=[(configname, 'Random sampling')],\n",
    "                        pvalues=[pvalue],\n",
    "                        pvalue_thresholds=[[1e-3, '**'], [1e-2, '*'],\n",
    "                                           [0.05, '*'], [1, 'ns']],\n",
    "                        test=None, text_format='star', loc='outside',\n",
    "                        line_height=0.02, linewidth=2.5, fontsize=24, verbose=0)\n",
    "    ax.set_xlabel('')\n",
    "    ax.set_ylabel(f\"{score}\", fontsize=20)\n",
    "    ax.tick_params(axis=\"x\", labelsize=20)\n",
    "    ax.tick_params(axis=\"y\", labelsize=14)\n",
    "    plt.suptitle(f\"{title}\", fontsize=30)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(f\"{save_directory}/images/pdf/{filename}_distribution.pdf\")\n",
    "    plt.savefig(f\"{save_directory}/images/png/{filename}_distribution.png\", bbox_inches='tight')\n",
    "\n",
    "\n",
    "def wrap_compute_to_plot(best_medians, FE_per_gen):\n",
    "    def compute_to_plot(medians, FE):\n",
    "        to_plot = []\n",
    "        xs = []\n",
    "        x = 0\n",
    "        for m in medians:\n",
    "            to_plot.append(m)\n",
    "            xs.append(x)\n",
    "            to_plot.append(m)\n",
    "            x += FE\n",
    "            xs.append(x)\n",
    "        return to_plot, xs\n",
    "\n",
    "    to_plot, xs = compute_to_plot(best_medians, FE_per_gen)\n",
    "    return to_plot, xs\n",
    "\n",
    "\n",
    "def plot_convergence(ev_best_medians, nev_best_median, FE_per_gen, score, configname, title, save_directory, filename):\n",
    "    to_plot_nev = [nev_best_median, nev_best_median]\n",
    "    to_plot_ev, xs = wrap_compute_to_plot(ev_best_medians, FE_per_gen)\n",
    "\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    plt.plot([xs[0], xs[-1]], to_plot_nev, color='orange', lw=2.5, label='Random sampling')\n",
    "    plt.plot(xs, to_plot_ev, color='blue', lw=2.5, label=configname)\n",
    "    plt.xlabel(\"Pipeline evaluation\", fontsize=20)\n",
    "    plt.ylabel(f\"{score}\", fontsize=20)\n",
    "    plt.title(f\"{title}\", fontsize=30)\n",
    "    plt.xticks(fontsize=14)\n",
    "    plt.yticks(fontsize=14)\n",
    "    plt.xlim(0, xs[-1])\n",
    "    plt.legend(fontsize=18)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(f\"{save_directory}/images/pdf/{filename}_converegence.pdf\")\n",
    "    plt.savefig(f\"{save_directory}/images/png/{filename}_convergence.png\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/scipy/stats/morestats.py:3141: UserWarning: Exact p-value calculation does not work if there are ties. Switching to normal approximation.\n",
      "  warnings.warn(\"Exact p-value calculation does not work if there are \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wilcoxon Two-sided pvalue: 0.0014988411849755196\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m/Users/matsumoton/Git/tpot_benchmarking/benchmarking_result.ipynb Cell 5'\u001b[0m in \u001b[0;36m<cell line: 7>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/matsumoton/Git/tpot_benchmarking/benchmarking_result.ipynb#ch0000005?line=13'>14</a>\u001b[0m \u001b[39mfor\u001b[39;00m i \u001b[39min\u001b[39;00m \u001b[39mrange\u001b[39m(numberofruns):\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/matsumoton/Git/tpot_benchmarking/benchmarking_result.ipynb#ch0000005?line=14'>15</a>\u001b[0m     ev_df_name \u001b[39m=\u001b[39m \u001b[39mf\u001b[39m\u001b[39m\"\u001b[39m\u001b[39m{\u001b[39;00mdirectoryev\u001b[39m}\u001b[39;00m\u001b[39m{\u001b[39;00minputname\u001b[39m}\u001b[39;00m\u001b[39m_run_\u001b[39m\u001b[39m{\u001b[39;00mi\u001b[39m}\u001b[39;00m\u001b[39m_evolution_pop20_gen10.csv\u001b[39m\u001b[39m\"\u001b[39m\n\u001b[0;32m---> <a href='vscode-notebook-cell:/Users/matsumoton/Git/tpot_benchmarking/benchmarking_result.ipynb#ch0000005?line=15'>16</a>\u001b[0m     ev_dfs\u001b[39m.\u001b[39mappend(pd\u001b[39m.\u001b[39;49mread_csv(ev_df_name, sep\u001b[39m=\u001b[39;49m\u001b[39m'\u001b[39;49m\u001b[39m,\u001b[39;49m\u001b[39m'\u001b[39;49m))\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/matsumoton/Git/tpot_benchmarking/benchmarking_result.ipynb#ch0000005?line=16'>17</a>\u001b[0m     nev_df_name \u001b[39m=\u001b[39m \u001b[39mf\u001b[39m\u001b[39m\"\u001b[39m\u001b[39m{\u001b[39;00mdirectoryev\u001b[39m}\u001b[39;00m\u001b[39m{\u001b[39;00minputname\u001b[39m}\u001b[39;00m\u001b[39m_run_\u001b[39m\u001b[39m{\u001b[39;00mi\u001b[39m}\u001b[39;00m\u001b[39m_no_evolution_pop100_gen0.csv\u001b[39m\u001b[39m\"\u001b[39m\n\u001b[1;32m     <a href='vscode-notebook-cell:/Users/matsumoton/Git/tpot_benchmarking/benchmarking_result.ipynb#ch0000005?line=17'>18</a>\u001b[0m     nev_dfs\u001b[39m.\u001b[39mappend(pd\u001b[39m.\u001b[39mread_csv(nev_df_name, sep\u001b[39m=\u001b[39m\u001b[39m'\u001b[39m\u001b[39m,\u001b[39m\u001b[39m'\u001b[39m))\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py:311\u001b[0m, in \u001b[0;36mdeprecate_nonkeyword_arguments.<locals>.decorate.<locals>.wrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py?line=304'>305</a>\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mlen\u001b[39m(args) \u001b[39m>\u001b[39m num_allow_args:\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py?line=305'>306</a>\u001b[0m     warnings\u001b[39m.\u001b[39mwarn(\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py?line=306'>307</a>\u001b[0m         msg\u001b[39m.\u001b[39mformat(arguments\u001b[39m=\u001b[39marguments),\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py?line=307'>308</a>\u001b[0m         \u001b[39mFutureWarning\u001b[39;00m,\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py?line=308'>309</a>\u001b[0m         stacklevel\u001b[39m=\u001b[39mstacklevel,\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py?line=309'>310</a>\u001b[0m     )\n\u001b[0;32m--> <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/util/_decorators.py?line=310'>311</a>\u001b[0m \u001b[39mreturn\u001b[39;00m func(\u001b[39m*\u001b[39;49margs, \u001b[39m*\u001b[39;49m\u001b[39m*\u001b[39;49mkwargs)\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py:680\u001b[0m, in \u001b[0;36mread_csv\u001b[0;34m(filepath_or_buffer, sep, delimiter, header, names, index_col, usecols, squeeze, prefix, mangle_dupe_cols, dtype, engine, converters, true_values, false_values, skipinitialspace, skiprows, skipfooter, nrows, na_values, keep_default_na, na_filter, verbose, skip_blank_lines, parse_dates, infer_datetime_format, keep_date_col, date_parser, dayfirst, cache_dates, iterator, chunksize, compression, thousands, decimal, lineterminator, quotechar, quoting, doublequote, escapechar, comment, encoding, encoding_errors, dialect, error_bad_lines, warn_bad_lines, on_bad_lines, delim_whitespace, low_memory, memory_map, float_precision, storage_options)\u001b[0m\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=664'>665</a>\u001b[0m kwds_defaults \u001b[39m=\u001b[39m _refine_defaults_read(\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=665'>666</a>\u001b[0m     dialect,\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=666'>667</a>\u001b[0m     delimiter,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=675'>676</a>\u001b[0m     defaults\u001b[39m=\u001b[39m{\u001b[39m\"\u001b[39m\u001b[39mdelimiter\u001b[39m\u001b[39m\"\u001b[39m: \u001b[39m\"\u001b[39m\u001b[39m,\u001b[39m\u001b[39m\"\u001b[39m},\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=676'>677</a>\u001b[0m )\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=677'>678</a>\u001b[0m kwds\u001b[39m.\u001b[39mupdate(kwds_defaults)\n\u001b[0;32m--> <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=679'>680</a>\u001b[0m \u001b[39mreturn\u001b[39;00m _read(filepath_or_buffer, kwds)\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py:581\u001b[0m, in \u001b[0;36m_read\u001b[0;34m(filepath_or_buffer, kwds)\u001b[0m\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=577'>578</a>\u001b[0m     \u001b[39mreturn\u001b[39;00m parser\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=579'>580</a>\u001b[0m \u001b[39mwith\u001b[39;00m parser:\n\u001b[0;32m--> <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=580'>581</a>\u001b[0m     \u001b[39mreturn\u001b[39;00m parser\u001b[39m.\u001b[39;49mread(nrows)\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py:1254\u001b[0m, in \u001b[0;36mTextFileReader.read\u001b[0;34m(self, nrows)\u001b[0m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=1251'>1252</a>\u001b[0m nrows \u001b[39m=\u001b[39m validate_integer(\u001b[39m\"\u001b[39m\u001b[39mnrows\u001b[39m\u001b[39m\"\u001b[39m, nrows)\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=1252'>1253</a>\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[0;32m-> <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=1253'>1254</a>\u001b[0m     index, columns, col_dict \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49m_engine\u001b[39m.\u001b[39;49mread(nrows)\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=1254'>1255</a>\u001b[0m \u001b[39mexcept\u001b[39;00m \u001b[39mException\u001b[39;00m:\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/readers.py?line=1255'>1256</a>\u001b[0m     \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mclose()\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/c_parser_wrapper.py:225\u001b[0m, in \u001b[0;36mCParserWrapper.read\u001b[0;34m(self, nrows)\u001b[0m\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/c_parser_wrapper.py?line=222'>223</a>\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/c_parser_wrapper.py?line=223'>224</a>\u001b[0m     \u001b[39mif\u001b[39;00m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mlow_memory:\n\u001b[0;32m--> <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/c_parser_wrapper.py?line=224'>225</a>\u001b[0m         chunks \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49m_reader\u001b[39m.\u001b[39;49mread_low_memory(nrows)\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/c_parser_wrapper.py?line=225'>226</a>\u001b[0m         \u001b[39m# destructive to chunks\u001b[39;00m\n\u001b[1;32m    <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/io/parsers/c_parser_wrapper.py?line=226'>227</a>\u001b[0m         data \u001b[39m=\u001b[39m _concatenate_chunks(chunks)\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/_libs/parsers.pyx:805\u001b[0m, in \u001b[0;36mpandas._libs.parsers.TextReader.read_low_memory\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/_libs/parsers.pyx:883\u001b[0m, in \u001b[0;36mpandas._libs.parsers.TextReader._read_rows\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/_libs/parsers.pyx:1026\u001b[0m, in \u001b[0;36mpandas._libs.parsers.TextReader._convert_column_data\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/_libs/parsers.pyx:1072\u001b[0m, in \u001b[0;36mpandas._libs.parsers.TextReader._convert_tokens\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/_libs/parsers.pyx:1147\u001b[0m, in \u001b[0;36mpandas._libs.parsers.TextReader._convert_with_dtype\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py:1429\u001b[0m, in \u001b[0;36mis_extension_array_dtype\u001b[0;34m(arr_or_dtype)\u001b[0m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1419'>1420</a>\u001b[0m     \u001b[39m# Note: if other EA dtypes are ever held in HybridBlock, exclude those\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1420'>1421</a>\u001b[0m     \u001b[39m#  here too.\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1421'>1422</a>\u001b[0m     \u001b[39m# NB: need to check DatetimeTZDtype and not is_datetime64tz_dtype\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1422'>1423</a>\u001b[0m     \u001b[39m#  to exclude ArrowTimestampUSDtype\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1423'>1424</a>\u001b[0m     \u001b[39mreturn\u001b[39;00m \u001b[39misinstance\u001b[39m(dtype, ExtensionDtype) \u001b[39mand\u001b[39;00m \u001b[39mnot\u001b[39;00m \u001b[39misinstance\u001b[39m(\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1424'>1425</a>\u001b[0m         dtype, (DatetimeTZDtype, PeriodDtype)\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1425'>1426</a>\u001b[0m     )\n\u001b[0;32m-> <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1428'>1429</a>\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39mis_extension_array_dtype\u001b[39m(arr_or_dtype) \u001b[39m-\u001b[39m\u001b[39m>\u001b[39m \u001b[39mbool\u001b[39m:\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1429'>1430</a>\u001b[0m     \u001b[39m\"\"\"\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1430'>1431</a>\u001b[0m \u001b[39m    Check if an object is a pandas extension array type.\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1431'>1432</a>\u001b[0m \n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1471'>1472</a>\u001b[0m \u001b[39m    False\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1472'>1473</a>\u001b[0m \u001b[39m    \"\"\"\u001b[39;00m\n\u001b[1;32m   <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/pandas/core/dtypes/common.py?line=1473'>1474</a>\u001b[0m     dtype \u001b[39m=\u001b[39m \u001b[39mgetattr\u001b[39m(arr_or_dtype, \u001b[39m\"\u001b[39m\u001b[39mdtype\u001b[39m\u001b[39m\"\u001b[39m, arr_or_dtype)\n",
      "File \u001b[0;32m_pydevd_bundle/pydevd_cython.pyx:1366\u001b[0m, in \u001b[0;36m_pydevd_bundle.pydevd_cython.SafeCallWrapper.__call__\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32m_pydevd_bundle/pydevd_cython.pyx:1667\u001b[0m, in \u001b[0;36m_pydevd_bundle.pydevd_cython.ThreadTracer.__call__\u001b[0;34m()\u001b[0m\n",
      "File \u001b[0;32m~/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/debugpy/_vendored/pydevd/_pydev_bundle/pydev_is_thread_alive.py:9\u001b[0m, in \u001b[0;36mis_thread_alive\u001b[0;34m(t)\u001b[0m\n\u001b[1;32m      <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/debugpy/_vendored/pydevd/_pydev_bundle/pydev_is_thread_alive.py?line=5'>6</a>\u001b[0m _temp \u001b[39m=\u001b[39m threading\u001b[39m.\u001b[39mThread()\n\u001b[1;32m      <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/debugpy/_vendored/pydevd/_pydev_bundle/pydev_is_thread_alive.py?line=6'>7</a>\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mhasattr\u001b[39m(_temp, \u001b[39m'\u001b[39m\u001b[39m_is_stopped\u001b[39m\u001b[39m'\u001b[39m):  \u001b[39m# Python 3.x has this\u001b[39;00m\n\u001b[0;32m----> <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/debugpy/_vendored/pydevd/_pydev_bundle/pydev_is_thread_alive.py?line=8'>9</a>\u001b[0m     \u001b[39mdef\u001b[39;00m \u001b[39mis_thread_alive\u001b[39m(t):\n\u001b[1;32m     <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/debugpy/_vendored/pydevd/_pydev_bundle/pydev_is_thread_alive.py?line=9'>10</a>\u001b[0m         \u001b[39mreturn\u001b[39;00m \u001b[39mnot\u001b[39;00m t\u001b[39m.\u001b[39m_is_stopped\n\u001b[1;32m     <a href='file:///Users/matsumoton/opt/anaconda3/envs/tpot_env/lib/python3.10/site-packages/debugpy/_vendored/pydevd/_pydev_bundle/pydev_is_thread_alive.py?line=11'>12</a>\u001b[0m \u001b[39melif\u001b[39;00m \u001b[39mhasattr\u001b[39m(_temp, \u001b[39m'\u001b[39m\u001b[39m_Thread__stopped\u001b[39m\u001b[39m'\u001b[39m):  \u001b[39m# Python 2.x has this\u001b[39;00m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "directoryev = \"/Users/matsumoton/Git/results_baseline/gen_fitnesses/\"\n",
    "ev_dfs = []\n",
    "nev_dfs = []\n",
    "\n",
    "score = 'Accuracy'\n",
    "\n",
    "for j in range(1,41):\n",
    "    ev_dfs = []\n",
    "    nev_dfs = []\n",
    "    inputname = \"digen\" + str(j)\n",
    "    configname = inputname + \"config\"\n",
    "    title = inputname\n",
    "    filename = inputname\n",
    "    for i in range(numberofruns):\n",
    "        ev_df_name = f\"{directoryev}{inputname}_run_{i}_evolution_pop20_gen10.csv\"\n",
    "        ev_dfs.append(pd.read_csv(ev_df_name, sep=','))\n",
    "        nev_df_name = f\"{directoryev}{inputname}_run_{i}_no_evolution_pop100_gen0.csv\"\n",
    "        nev_dfs.append(pd.read_csv(nev_df_name, sep=','))\n",
    "\n",
    "    try:\n",
    "        makedirs(f\"{directoryev}images/pdf/\")\n",
    "        makedirs(f\"{directoryev}images/png/\")\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "    ev_bests, nev_bests = extract_bests(ev_dfs, nev_dfs)\n",
    "    ev_best_medians, nev_best_median = compute_medians_bests(ev_bests, nev_bests)\n",
    "\n",
    "    last_gen_bests = [np.max(c) for c in ev_bests.T]\n",
    "    random_bests = [x for x in nev_bests]\n",
    "    plot_boxplots(last_gen_bests, random_bests, score, configname, title, directoryev, filename)\n",
    "\n",
    "    FE_per_gen = 10\n",
    "\n",
    "    plot_convergence(ev_best_medians, nev_best_median, FE_per_gen, score, configname, title, directoryev,\n",
    "                        filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/Users/matsumoton/Git/tpot_benchmarking\n",
      "2.53125\n",
      "This is dataset 'AP_Endometrium_Uterus', the target feature is 'Tissue'\n",
      "URL: https://api.openml.org/data/v1/download/54047/AP_Endometrium_Uterus.arff\n",
      "**Author**:   \n",
      "**Source**: Unknown - Date unknown  \n",
      "**Please cite**:   \n",
      "\n",
      "GEMLeR provides a collection of gene expression datasets that can be used for benchmarking gene expression oriented machine learning algorithms. They can be used for estimation of different quality metrics (e.g. accuracy, precision, area under ROC curve, etc.) for classification, feature selection or clustering algorithms.\n",
      "\n",
      "This repository was inspired by an increasing need in machine learning / bioinformatics communities f\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "((148, 10935), (37, 10935), (148,), (37,))"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "os.chdir('/Users/matsumoton/Git/tpot_benchmarking')\n",
    "print(os.getcwd())\n",
    "from tpot.tpot import TPOTClassifier\n",
    "from sklearn.datasets import load_digits\n",
    "from sklearn.model_selection import train_test_split\n",
    "from test_utils import extract_labels, get_optimizer, create_dirs\n",
    "\n",
    "\n",
    "import openml\n",
    "import pandas as pd\n",
    "from openml.datasets import edit_dataset, fork_dataset, get_dataset\n",
    " # This is done based on the dataset ID.\n",
    "dataset = openml.datasets.get_dataset(1164)\n",
    "\n",
    "# Print a summary\n",
    "print(\n",
    "    f\"This is dataset '{dataset.name}', the target feature is \"\n",
    "    f\"'{dataset.default_target_attribute}'\"\n",
    ")\n",
    "print(f\"URL: {dataset.url}\")\n",
    "print(dataset.description[:500])\n",
    "\n",
    "X, y, categorical_indicator, attribute_names = dataset.get_data(\n",
    "    dataset_format=\"array\", target=dataset.default_target_attribute\n",
    ")\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.8,random_state=5)\n",
    "\n",
    "#digits = load_digits()\n",
    "#X_train, X_test, y_train, y_test = train_test_split(digits.data, digits.target,train_size=0.75, test_size=0.25)\n",
    "X_train.shape, X_test.shape, y_train.shape, y_test.shape\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import sklearn.decomposition\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "# Instantiate PCA\n",
    "#\n",
    "pca = decomposition.PCA()\n",
    "#\n",
    "# Determine transformed features\n",
    "#\n",
    "X_train_pca = pca.fit_transform(X_train_std)\n",
    "#\n",
    "# Determine explained variance using explained_variance_ration_ attribute\n",
    "#\n",
    "exp_var_pca = pca.explained_variance_ratio_\n",
    "#\n",
    "# Cumulative sum of eigenvalues; This will be used to create step plot\n",
    "# for visualizing the variance explained by each principal component.\n",
    "#\n",
    "cum_sum_eigenvalues = np.cumsum(exp_var_pca)\n",
    "#\n",
    "# Create the visualization plot\n",
    "#\n",
    "plt.bar(range(0,len(exp_var_pca)), exp_var_pca, alpha=0.5, align='center', label='Individual explained variance')\n",
    "plt.step(range(0,len(cum_sum_eigenvalues)), cum_sum_eigenvalues, where='mid',label='Cumulative explained variance')\n",
    "plt.ylabel('Explained variance ratio')\n",
    "plt.xlabel('Principal component index')\n",
    "plt.legend(loc='best')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([8.8210359e-02, 7.0886120e-02, 5.1611274e-02, 4.4747870e-02,\n",
       "       3.1382047e-02, 2.9136149e-02, 2.5107743e-02, 2.4355579e-02,\n",
       "       2.3528580e-02, 2.1284571e-02, 2.0286333e-02, 1.9795828e-02,\n",
       "       1.8350577e-02, 1.6577961e-02, 1.5768850e-02, 1.3979936e-02,\n",
       "       1.3856016e-02, 1.3054941e-02, 1.2421572e-02, 1.1780609e-02,\n",
       "       1.1245822e-02, 1.0441891e-02, 1.0074229e-02, 9.8397769e-03,\n",
       "       9.7227786e-03, 9.1858897e-03, 8.7880250e-03, 8.5602105e-03,\n",
       "       8.2416143e-03, 7.9613570e-03, 7.8919744e-03, 7.4135554e-03,\n",
       "       7.1611498e-03, 6.9767623e-03, 6.7103985e-03, 6.5905056e-03,\n",
       "       6.5012504e-03, 6.3884496e-03, 6.1244732e-03, 6.0603423e-03,\n",
       "       5.9416476e-03, 5.7772403e-03, 5.6067775e-03, 5.5156746e-03,\n",
       "       5.4597845e-03, 5.3077675e-03, 5.0373739e-03, 4.9276571e-03,\n",
       "       4.9007279e-03, 4.7832145e-03, 4.7459882e-03, 4.6546012e-03,\n",
       "       4.5432299e-03, 4.4547971e-03, 4.3643853e-03, 4.2720260e-03,\n",
       "       4.2079096e-03, 4.1392446e-03, 4.0791216e-03, 4.0308950e-03,\n",
       "       3.9594038e-03, 3.8779355e-03, 3.8074141e-03, 3.6969120e-03,\n",
       "       3.5787297e-03, 3.5505111e-03, 3.5370681e-03, 3.4272806e-03,\n",
       "       3.3720522e-03, 3.3412350e-03, 3.2974568e-03, 3.2202362e-03,\n",
       "       3.1726216e-03, 3.0818703e-03, 3.0735538e-03, 3.0024068e-03,\n",
       "       2.9340894e-03, 2.8909435e-03, 2.8386272e-03, 2.7806978e-03,\n",
       "       2.7393408e-03, 2.6866884e-03, 2.6370327e-03, 2.5866374e-03,\n",
       "       2.5724811e-03, 2.5543196e-03, 2.4944639e-03, 2.4264178e-03,\n",
       "       2.4066605e-03, 2.3478745e-03, 2.3345703e-03, 2.3107983e-03,\n",
       "       2.2749042e-03, 2.2231413e-03, 2.2007716e-03, 2.1312952e-03,\n",
       "       2.0722144e-03, 2.0718903e-03, 2.0205062e-03, 1.9831490e-03,\n",
       "       1.9462398e-03, 1.9388049e-03, 1.8956874e-03, 1.8501997e-03,\n",
       "       1.8155444e-03, 1.7910379e-03, 1.7653462e-03, 1.7193764e-03,\n",
       "       1.7125020e-03, 1.6705224e-03, 1.6581218e-03, 1.6314060e-03,\n",
       "       1.6188156e-03, 1.5967162e-03, 1.5472384e-03, 1.5319396e-03,\n",
       "       1.4942519e-03, 1.4811504e-03, 1.4619685e-03, 1.4279517e-03,\n",
       "       1.3986119e-03, 1.3739845e-03, 1.3522084e-03, 1.3150326e-03,\n",
       "       1.3118614e-03, 1.2918527e-03, 1.2584317e-03, 1.2354931e-03,\n",
       "       1.1856278e-03, 1.1773405e-03, 1.1494068e-03, 1.1138656e-03,\n",
       "       1.0763044e-03, 1.0208376e-03, 9.8257046e-04, 9.6179650e-04,\n",
       "       9.3114114e-04, 9.0833864e-04, 8.8530063e-04, 8.6374412e-04,\n",
       "       8.5429323e-04, 8.2147686e-04, 7.9433242e-04, 7.7237893e-04,\n",
       "       7.4156007e-04, 7.0924108e-04, 6.8852695e-04, 1.2684094e-15],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exp_var_pca"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "8903c8b9fec32094f84829a494fa74dc8559ee325273db181526c71a335012d6"
  },
  "kernelspec": {
   "display_name": "Python 3.10.4 ('tpot_env')",
   "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": "3.10.4"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}