docs/en_US/source/examples/63.MOEAD_example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MOEA/D Searcher example\n",
    "\n",
    "This is an example about how using MOEADSearcher for multi-objectives optimization."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Import modules and prepare data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from hypernets.core.random_state import set_random_state\n",
    "set_random_state(1234)\n",
    "\n",
    "from hypernets.utils import logging as hyn_logging\n",
    "from hypernets.examples.plain_model import PlainModel, PlainSearchSpace\n",
    "from hypernets.searchers.nsga_searcher import RNSGAIISearcher\n",
    "\n",
    "from hypergbm import make_experiment\n",
    "\n",
    "from hypernets.tabular import get_tool_box\n",
    "from hypernets.tabular.datasets import dsutils\n",
    "from hypernets.tabular.sklearn_ex import MultiLabelEncoder\n",
    "\n",
    "\n",
    "hyn_logging.set_level(hyn_logging.WARN)\n",
    "\n",
    "df = dsutils.load_bank().head(1000)\n",
    "tb = get_tool_box(df)\n",
    "df_train, df_test = tb.train_test_split(df, test_size=0.2, random_state=9527)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Run an experiment within NSGAIISearcher"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MOEADSearcher(objectives=[PredictionObjective(name=logloss, scorer=make_scorer(log_loss, needs_proba=True), direction=min), NumOfFeatures(name=nf, sample_size=1000, direction=min)], n_neighbors=2, recombination=SinglePointCrossOver(random_state=RandomState(MT19937)), mutation=SinglePointMutation(random_state=RandomState(MT19937), proba=0.7), population_size=6)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "experiment = make_experiment(df_train,\n",
    "                             eval_data=df_test.copy(),\n",
    "                             callbacks=[],\n",
    "                             random_state=1234,\n",
    "                             search_callbacks=[],\n",
    "                             target='y',\n",
    "                             searcher='moead',  # available MOO searcher: moead, nsga2, rnsga2\n",
    "                             reward_metric='logloss',\n",
    "                             objectives=['nf'],\n",
    "                             drift_detection=False,\n",
    "                             early_stopping_rounds=30)\n",
    "\n",
    "estimators = experiment.run(max_trials=30)\n",
    "hyper_model = experiment.hyper_model_\n",
    "hyper_model.searcher"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Summary trails"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>trial_no</th>\n",
       "      <th>succeeded</th>\n",
       "      <th>elapsed</th>\n",
       "      <th>non_dominated</th>\n",
       "      <th>model_index</th>\n",
       "      <th>reward_logloss</th>\n",
       "      <th>reward_nf</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>True</td>\n",
       "      <td>0.446323</td>\n",
       "      <td>True</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.217409</td>\n",
       "      <td>0.625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>True</td>\n",
       "      <td>4.100305</td>\n",
       "      <td>True</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.537368</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>True</td>\n",
       "      <td>4.796208</td>\n",
       "      <td>True</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.253515</td>\n",
       "      <td>0.125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>True</td>\n",
       "      <td>1.060251</td>\n",
       "      <td>True</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.246395</td>\n",
       "      <td>0.5625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>30</td>\n",
       "      <td>True</td>\n",
       "      <td>0.366623</td>\n",
       "      <td>True</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.177716</td>\n",
       "      <td>0.75</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    trial_no  succeeded   elapsed  non_dominated  model_index reward_logloss  \\\n",
       "4          5       True  0.446323           True          0.0       0.217409   \n",
       "6          7       True  4.100305           True          1.0       0.537368   \n",
       "8          9       True  4.796208           True          2.0       0.253515   \n",
       "9         10       True  1.060251           True          3.0       0.246395   \n",
       "22        30       True  0.366623           True          4.0       0.177716   \n",
       "\n",
       "   reward_nf  \n",
       "4      0.625  \n",
       "6        0.0  \n",
       "8      0.125  \n",
       "9     0.5625  \n",
       "22      0.75  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_trials = hyper_model.history.to_df().copy().drop(['scores', 'reward'], axis=1)\n",
    "df_trials[df_trials['non_dominated'] == True]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Plot pareto font\n",
    "\n",
    "We can pick model accord to Decision Maker's preferences from the pareto plot, the number in the figure indicates the index of pipeline models.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax  = hyper_model.history.plot_best_trials()\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Plot population"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax  = hyper_model.searcher.plot_population()\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Evaluate the selected model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of pipeline: 5 \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'auc': 0.8417038690476191,\n",
       " 'accuracy': 0.855,\n",
       " 'f1': 0.17142857142857143,\n",
       " 'recall': 0.09375,\n",
       " 'precision': 1.0}"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(f\"Number of pipeline: {len(estimators)} \")\n",
    "\n",
    "pipeline_model = estimators[0]  # selection the first pipeline model\n",
    "X_test = df_test.copy()\n",
    "y_test = X_test.pop('y')\n",
    "\n",
    "preds = pipeline_model.predict(X_test)\n",
    "proba = pipeline_model.predict_proba(X_test)\n",
    "\n",
    "tb.metrics.calc_score(y_test, preds, proba, metrics=['auc', 'accuracy', 'f1', 'recall', 'precision'], pos_label=\"yes\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Automatically convert metric to negatives for minimize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>trial_no</th>\n",
       "      <th>succeeded</th>\n",
       "      <th>elapsed</th>\n",
       "      <th>non_dominated</th>\n",
       "      <th>model_index</th>\n",
       "      <th>reward_accuracy</th>\n",
       "      <th>reward_precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>True</td>\n",
       "      <td>0.645476</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.91</td>\n",
       "      <td>-0.777778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>True</td>\n",
       "      <td>0.752835</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.8975</td>\n",
       "      <td>-0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>True</td>\n",
       "      <td>0.779298</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.8975</td>\n",
       "      <td>-0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>True</td>\n",
       "      <td>0.737511</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.905</td>\n",
       "      <td>-0.625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>True</td>\n",
       "      <td>0.498275</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.90625</td>\n",
       "      <td>-0.733333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   trial_no  succeeded   elapsed  non_dominated  model_index reward_accuracy  \\\n",
       "0         1       True  0.645476          False          NaN           -0.91   \n",
       "1         2       True  0.752835          False          NaN         -0.8975   \n",
       "2         3       True  0.779298          False          NaN         -0.8975   \n",
       "3         4       True  0.737511          False          NaN          -0.905   \n",
       "4         5       True  0.498275          False          NaN        -0.90625   \n",
       "\n",
       "  reward_precision  \n",
       "0        -0.777778  \n",
       "1             -0.0  \n",
       "2             -0.0  \n",
       "3           -0.625  \n",
       "4        -0.733333  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "experiment = make_experiment(df_train,\n",
    "                             eval_data=df_test.copy(),\n",
    "                             callbacks=[],\n",
    "                             random_state=1234,\n",
    "                             search_callbacks=[],\n",
    "                             target='y',\n",
    "                             pos_label=\"yes\",\n",
    "                             searcher='moead',\n",
    "                             reward_metric='accuracy',\n",
    "                             objectives=['precision'],\n",
    "                             drift_detection=False,\n",
    "                             early_stopping_rounds=30)\n",
    "\n",
    "estimators = experiment.run(max_trials=30)\n",
    "hyper_model = experiment.hyper_model_\n",
    "hyper_model.history.to_df().copy().drop(['scores', 'reward'], axis=1)[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.8.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}