{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### This notebook shows how you can look under the hood of the PyCaret Time Series Module and customize the AutoML flow per your business needs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pprint import pprint\n",
    "import pandas as pd\n",
    "\n",
    "from pycaret.datasets import get_data\n",
    "from pycaret.internal.pycaret_experiment import TimeSeriesExperiment\n",
    "\n",
    "from sktime.utils.plotting import plot_series"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y = get_data('airline', verbose=False)\n",
    "_ = plot_series(y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup PyCaret Time Series Experiment \n",
    "\n",
    "* Based on business needs and good data science principles\n",
    "  - e.g. Interested in forecasting 12 months out\n",
    "  - Use 3 folds to do any cross-validation\n",
    "  \n",
    "* **Important arguments to setup**\n",
    "  - `fh` - forecast-horizon\n",
    "  - `folds` - number of cross validation folds to use\n",
    "  - `fold_strategy` - 'sliding' or 'expanding'\n",
    "  - `seasonal_period`: Inferred using index; can be explicitly set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "</style><table id=\"T_2a54f_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Description</th>        <th class=\"col_heading level0 col1\" >Value</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_2a54f_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_2a54f_row0_col0\" class=\"data row0 col0\" >session_id</td>\n",
       "                        <td id=\"T_2a54f_row0_col1\" class=\"data row0 col1\" >42</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_2a54f_row1_col0\" class=\"data row1 col0\" >Original Data</td>\n",
       "                        <td id=\"T_2a54f_row1_col1\" class=\"data row1 col1\" >(144, 1)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_2a54f_row2_col0\" class=\"data row2 col0\" >Missing Values</td>\n",
       "                        <td id=\"T_2a54f_row2_col1\" class=\"data row2 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_2a54f_row3_col0\" class=\"data row3 col0\" >Transformed Train Set</td>\n",
       "                        <td id=\"T_2a54f_row3_col1\" class=\"data row3 col1\" >(132,)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_2a54f_row4_col0\" class=\"data row4 col0\" >Transformed Test Set</td>\n",
       "                        <td id=\"T_2a54f_row4_col1\" class=\"data row4 col1\" >(12,)</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row5\" class=\"row_heading level0 row5\" >5</th>\n",
       "                        <td id=\"T_2a54f_row5_col0\" class=\"data row5 col0\" >Fold Generator</td>\n",
       "                        <td id=\"T_2a54f_row5_col1\" class=\"data row5 col1\" >ExpandingWindowSplitter</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row6\" class=\"row_heading level0 row6\" >6</th>\n",
       "                        <td id=\"T_2a54f_row6_col0\" class=\"data row6 col0\" >Fold Number</td>\n",
       "                        <td id=\"T_2a54f_row6_col1\" class=\"data row6 col1\" >3</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row7\" class=\"row_heading level0 row7\" >7</th>\n",
       "                        <td id=\"T_2a54f_row7_col0\" class=\"data row7 col0\" >CPU Jobs</td>\n",
       "                        <td id=\"T_2a54f_row7_col1\" class=\"data row7 col1\" >-1</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row8\" class=\"row_heading level0 row8\" >8</th>\n",
       "                        <td id=\"T_2a54f_row8_col0\" class=\"data row8 col0\" >Use GPU</td>\n",
       "                        <td id=\"T_2a54f_row8_col1\" class=\"data row8 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row9\" class=\"row_heading level0 row9\" >9</th>\n",
       "                        <td id=\"T_2a54f_row9_col0\" class=\"data row9 col0\" >Log Experiment</td>\n",
       "                        <td id=\"T_2a54f_row9_col1\" class=\"data row9 col1\" >False</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row10\" class=\"row_heading level0 row10\" >10</th>\n",
       "                        <td id=\"T_2a54f_row10_col0\" class=\"data row10 col0\" >Experiment Name</td>\n",
       "                        <td id=\"T_2a54f_row10_col1\" class=\"data row10 col1\" >ts-default-name</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row11\" class=\"row_heading level0 row11\" >11</th>\n",
       "                        <td id=\"T_2a54f_row11_col0\" class=\"data row11 col0\" >USI</td>\n",
       "                        <td id=\"T_2a54f_row11_col1\" class=\"data row11 col1\" >44d0</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_2a54f_level0_row12\" class=\"row_heading level0 row12\" >12</th>\n",
       "                        <td id=\"T_2a54f_row12_col0\" class=\"data row12 col0\" >Imputation Type</td>\n",
       "                        <td id=\"T_2a54f_row12_col1\" class=\"data row12 col1\" >simple</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e580b66408>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<pycaret.internal.pycaret_experiment.time_series_experiment.TimeSeriesExperiment at 0x1e5ffe33fc8>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exp = TimeSeriesExperiment()\n",
    "exp.setup(data=y, fh=12, session_id=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Looking under the hood\n",
    "\n",
    "### Train/Test Split\n",
    "\n",
    "Internally split: Keep len(fh) as test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_train = exp.get_config(\"y_train\")\n",
    "y_test = exp.get_config(\"y_test\")\n",
    "\n",
    "_ = plot_series(y_train, y_test, labels=['Train', 'Test'])\n",
    "_ = plot_series(y, y[-12:], labels=['Full Data', 'Last 12 points'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Seasonality\n",
    "\n",
    "Used to define the internal grid for tuning models, etc."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Seasonality Present: True\n",
      "Seasonal Period: 12\n"
     ]
    }
   ],
   "source": [
    "print(f\"\\nSeasonality Present: {exp.get_config('seasonality_present')}\")\n",
    "print(f\"Seasonal Period: {exp.get_config('seasonal_period')}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cross Validation / Fold Generation\n",
    "\n",
    "* Uses `sktime`'s `ExpandingWindowSplitter` and `SlidingWindowSplitter` to generate folds.\n",
    "* Arguments for the splitters are determined based on inputs to `exp.setup`\n",
    "\n",
    "**Example** (using `sktime=0.5.3`):\n",
    "  - Total Data Points = 144\n",
    "  - Train Data Points = 144 - 12 = 132\n",
    "  - Folds Requested = 3 (Default)\n",
    "  - Type = Expanding (Default)\n",
    "  - Initial Window = Computed (132 - 3 * 12 = 96)\n",
    "  - Step Size = Window Length = Length of Forecast Horizon = 12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Fold Generator: <sktime.forecasting.model_selection._split.ExpandingWindowSplitter object at 0x000001E5800A4688>\n",
      "  - Fold Generator Horizon: [ 1  2  3  4  5  6  7  8  9 10 11 12]\n",
      "  - Fold Generator Initial Window: 96\n",
      "  - Fold Generator Step Length: 12\n",
      "  - Fold Generator Window Length: 12\n"
     ]
    }
   ],
   "source": [
    "print(f\"\\nFold Generator: {exp.get_config('fold_generator')}\")\n",
    "print(f\"  - Fold Generator Horizon: {exp.get_config('fold_generator').fh}\")\n",
    "print(f\"  - Fold Generator Initial Window: {exp.get_config('fold_generator').initial_window}\")\n",
    "print(f\"  - Fold Generator Step Length: {exp.get_config('fold_generator').step_length}\")\n",
    "print(f\"  - Fold Generator Window Length: {exp.get_config('fold_generator').window_length}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Looking Under the Hood & Customizing the Flow"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Definitions\n",
    "\n",
    "- `Args` - Used in `create_model`\n",
    "- `Tune Grid` - Used for fixed grid search\n",
    "- `Tune Distributions` - Used in random grid search"
   ]
  },
  {
   "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>Name</th>\n",
       "      <th>Reference</th>\n",
       "      <th>Turbo</th>\n",
       "      <th>Special</th>\n",
       "      <th>Class</th>\n",
       "      <th>Equality</th>\n",
       "      <th>Args</th>\n",
       "      <th>Tune Grid</th>\n",
       "      <th>Tune Distributions</th>\n",
       "      <th>Tune Args</th>\n",
       "      <th>GPU Enabled</th>\n",
       "      <th>Tunable Class</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>naive</th>\n",
       "      <td>Naive Forecaster</td>\n",
       "      <td>sktime.forecasting.naive.NaiveForecaster</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>&lt;class 'sktime.forecasting.naive.NaiveForecast...</td>\n",
       "      <td>&lt;function ModelContainer.__init__.&lt;locals&gt;.&lt;la...</td>\n",
       "      <td>{}</td>\n",
       "      <td>{'strategy': ['last', 'mean', 'drift'], 'sp': ...</td>\n",
       "      <td>{'strategy': CategoricalDistribution(values=['...</td>\n",
       "      <td>{}</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>snaive</th>\n",
       "      <td>Seasonal Naive Forecaster</td>\n",
       "      <td>sktime.forecasting.naive.NaiveForecaster</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>&lt;class 'sktime.forecasting.naive.NaiveForecast...</td>\n",
       "      <td>&lt;function ModelContainer.__init__.&lt;locals&gt;.&lt;la...</td>\n",
       "      <td>{'sp': 12}</td>\n",
       "      <td>{'strategy': ['last', 'mean', 'drift'], 'sp': ...</td>\n",
       "      <td>{'strategy': CategoricalDistribution(values=['...</td>\n",
       "      <td>{}</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>polytrend</th>\n",
       "      <td>Polynomial Trend Forecaster</td>\n",
       "      <td>sktime.forecasting.trend.PolynomialTrendForeca...</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>&lt;class 'sktime.forecasting.trend.PolynomialTre...</td>\n",
       "      <td>&lt;function ModelContainer.__init__.&lt;locals&gt;.&lt;la...</td>\n",
       "      <td>{}</td>\n",
       "      <td>{'degree': [1, 2, 3, 4, 5], 'with_intercept': ...</td>\n",
       "      <td>{'degree': IntUniformDistribution(lower=1, upp...</td>\n",
       "      <td>{}</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>arima</th>\n",
       "      <td>ARIMA</td>\n",
       "      <td>sktime.forecasting.arima.ARIMA</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>&lt;class 'sktime.forecasting.arima.ARIMA'&gt;</td>\n",
       "      <td>&lt;function ModelContainer.__init__.&lt;locals&gt;.&lt;la...</td>\n",
       "      <td>{'seasonal_order': (0, 1, 0, 12), 'n_jobs': -1}</td>\n",
       "      <td>{'order': [(0, 0, 1), (0, 0, 0)], 'seasonal_or...</td>\n",
       "      <td>{'order': CategoricalDistribution(values=[(0, ...</td>\n",
       "      <td>{}</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>auto_arima</th>\n",
       "      <td>Auto ARIMA</td>\n",
       "      <td>sktime.forecasting.arima.AutoARIMA</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>&lt;class 'sktime.forecasting.arima.AutoARIMA'&gt;</td>\n",
       "      <td>&lt;function ModelContainer.__init__.&lt;locals&gt;.&lt;la...</td>\n",
       "      <td>{'sp': 12, 'random_state': 42, 'suppress_warni...</td>\n",
       "      <td>{'max_p': [12], 'max_q': [12], 'max_order': [N...</td>\n",
       "      <td>{'sp': CategoricalDistribution(values=[12, 24]...</td>\n",
       "      <td>{}</td>\n",
       "      <td>False</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   Name  \\\n",
       "ID                                        \n",
       "naive                  Naive Forecaster   \n",
       "snaive        Seasonal Naive Forecaster   \n",
       "polytrend   Polynomial Trend Forecaster   \n",
       "arima                             ARIMA   \n",
       "auto_arima                   Auto ARIMA   \n",
       "\n",
       "                                                    Reference  Turbo  Special  \\\n",
       "ID                                                                              \n",
       "naive                sktime.forecasting.naive.NaiveForecaster   True    False   \n",
       "snaive               sktime.forecasting.naive.NaiveForecaster   True    False   \n",
       "polytrend   sktime.forecasting.trend.PolynomialTrendForeca...   True    False   \n",
       "arima                          sktime.forecasting.arima.ARIMA   True    False   \n",
       "auto_arima                 sktime.forecasting.arima.AutoARIMA   True    False   \n",
       "\n",
       "                                                        Class  \\\n",
       "ID                                                              \n",
       "naive       <class 'sktime.forecasting.naive.NaiveForecast...   \n",
       "snaive      <class 'sktime.forecasting.naive.NaiveForecast...   \n",
       "polytrend   <class 'sktime.forecasting.trend.PolynomialTre...   \n",
       "arima                <class 'sktime.forecasting.arima.ARIMA'>   \n",
       "auto_arima       <class 'sktime.forecasting.arima.AutoARIMA'>   \n",
       "\n",
       "                                                     Equality  \\\n",
       "ID                                                              \n",
       "naive       <function ModelContainer.__init__.<locals>.<la...   \n",
       "snaive      <function ModelContainer.__init__.<locals>.<la...   \n",
       "polytrend   <function ModelContainer.__init__.<locals>.<la...   \n",
       "arima       <function ModelContainer.__init__.<locals>.<la...   \n",
       "auto_arima  <function ModelContainer.__init__.<locals>.<la...   \n",
       "\n",
       "                                                         Args  \\\n",
       "ID                                                              \n",
       "naive                                                      {}   \n",
       "snaive                                             {'sp': 12}   \n",
       "polytrend                                                  {}   \n",
       "arima         {'seasonal_order': (0, 1, 0, 12), 'n_jobs': -1}   \n",
       "auto_arima  {'sp': 12, 'random_state': 42, 'suppress_warni...   \n",
       "\n",
       "                                                    Tune Grid  \\\n",
       "ID                                                              \n",
       "naive       {'strategy': ['last', 'mean', 'drift'], 'sp': ...   \n",
       "snaive      {'strategy': ['last', 'mean', 'drift'], 'sp': ...   \n",
       "polytrend   {'degree': [1, 2, 3, 4, 5], 'with_intercept': ...   \n",
       "arima       {'order': [(0, 0, 1), (0, 0, 0)], 'seasonal_or...   \n",
       "auto_arima  {'max_p': [12], 'max_q': [12], 'max_order': [N...   \n",
       "\n",
       "                                           Tune Distributions Tune Args  \\\n",
       "ID                                                                        \n",
       "naive       {'strategy': CategoricalDistribution(values=['...        {}   \n",
       "snaive      {'strategy': CategoricalDistribution(values=['...        {}   \n",
       "polytrend   {'degree': IntUniformDistribution(lower=1, upp...        {}   \n",
       "arima       {'order': CategoricalDistribution(values=[(0, ...        {}   \n",
       "auto_arima  {'sp': CategoricalDistribution(values=[12, 24]...        {}   \n",
       "\n",
       "            GPU Enabled Tunable Class  \n",
       "ID                                     \n",
       "naive             False          None  \n",
       "snaive            False          None  \n",
       "polytrend         False          None  \n",
       "arima             False          None  \n",
       "auto_arima        False          None  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exp.models(internal=True).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Default"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_6c6ed_row3_col0,#T_6c6ed_row3_col1,#T_6c6ed_row3_col2,#T_6c6ed_row3_col3,#T_6c6ed_row3_col4,#T_6c6ed_row3_col5{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_6c6ed_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >cutoff</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_6c6ed_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_6c6ed_row0_col0\" class=\"data row0 col0\" >1956-12</td>\n",
       "                        <td id=\"T_6c6ed_row0_col1\" class=\"data row0 col1\" >63.4167</td>\n",
       "                        <td id=\"T_6c6ed_row0_col2\" class=\"data row0 col2\" >6967.9167</td>\n",
       "                        <td id=\"T_6c6ed_row0_col3\" class=\"data row0 col3\" >0.1549</td>\n",
       "                        <td id=\"T_6c6ed_row0_col4\" class=\"data row0 col4\" >0.1766</td>\n",
       "                        <td id=\"T_6c6ed_row0_col5\" class=\"data row0 col5\" >-1.2681</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_6c6ed_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_6c6ed_row1_col0\" class=\"data row1 col0\" >1957-12</td>\n",
       "                        <td id=\"T_6c6ed_row1_col1\" class=\"data row1 col1\" >52.3333</td>\n",
       "                        <td id=\"T_6c6ed_row1_col2\" class=\"data row1 col2\" >5842.1667</td>\n",
       "                        <td id=\"T_6c6ed_row1_col3\" class=\"data row1 col3\" >0.1205</td>\n",
       "                        <td id=\"T_6c6ed_row1_col4\" class=\"data row1 col4\" >0.1351</td>\n",
       "                        <td id=\"T_6c6ed_row1_col5\" class=\"data row1 col5\" >-0.5305</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_6c6ed_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_6c6ed_row2_col0\" class=\"data row2 col0\" >1958-12</td>\n",
       "                        <td id=\"T_6c6ed_row2_col1\" class=\"data row2 col1\" >91.3333</td>\n",
       "                        <td id=\"T_6c6ed_row2_col2\" class=\"data row2 col2\" >12811.6667</td>\n",
       "                        <td id=\"T_6c6ed_row2_col3\" class=\"data row2 col3\" >0.1954</td>\n",
       "                        <td id=\"T_6c6ed_row2_col4\" class=\"data row2 col4\" >0.2259</td>\n",
       "                        <td id=\"T_6c6ed_row2_col5\" class=\"data row2 col5\" >-1.8662</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_6c6ed_level0_row3\" class=\"row_heading level0 row3\" >Mean</th>\n",
       "                        <td id=\"T_6c6ed_row3_col0\" class=\"data row3 col0\" >nan</td>\n",
       "                        <td id=\"T_6c6ed_row3_col1\" class=\"data row3 col1\" >69.0278</td>\n",
       "                        <td id=\"T_6c6ed_row3_col2\" class=\"data row3 col2\" >8540.5833</td>\n",
       "                        <td id=\"T_6c6ed_row3_col3\" class=\"data row3 col3\" >0.1569</td>\n",
       "                        <td id=\"T_6c6ed_row3_col4\" class=\"data row3 col4\" >0.1792</td>\n",
       "                        <td id=\"T_6c6ed_row3_col5\" class=\"data row3 col5\" >-1.2216</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_6c6ed_level0_row4\" class=\"row_heading level0 row4\" >SD</th>\n",
       "                        <td id=\"T_6c6ed_row4_col0\" class=\"data row4 col0\" >nan</td>\n",
       "                        <td id=\"T_6c6ed_row4_col1\" class=\"data row4 col1\" >16.4086</td>\n",
       "                        <td id=\"T_6c6ed_row4_col2\" class=\"data row4 col2\" >3054.8806</td>\n",
       "                        <td id=\"T_6c6ed_row4_col3\" class=\"data row4 col3\" >0.0306</td>\n",
       "                        <td id=\"T_6c6ed_row4_col4\" class=\"data row4 col4\" >0.0371</td>\n",
       "                        <td id=\"T_6c6ed_row4_col5\" class=\"data row4 col5\" >0.5463</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e580cfb3c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "NaiveForecaster(sp=1, strategy='last', window_length=None)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = exp.create_model(\"naive\")\n",
    "model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Manually specify model arguments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_f54a4_row3_col0,#T_f54a4_row3_col1,#T_f54a4_row3_col2,#T_f54a4_row3_col3,#T_f54a4_row3_col4,#T_f54a4_row3_col5{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_f54a4_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >cutoff</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_f54a4_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_f54a4_row0_col0\" class=\"data row0 col0\" >1956-12</td>\n",
       "                        <td id=\"T_f54a4_row0_col1\" class=\"data row0 col1\" >54.5675</td>\n",
       "                        <td id=\"T_f54a4_row0_col2\" class=\"data row0 col2\" >5389.9425</td>\n",
       "                        <td id=\"T_f54a4_row0_col3\" class=\"data row0 col3\" >0.1332</td>\n",
       "                        <td id=\"T_f54a4_row0_col4\" class=\"data row0 col4\" >0.1489</td>\n",
       "                        <td id=\"T_f54a4_row0_col5\" class=\"data row0 col5\" >-0.7545</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_f54a4_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_f54a4_row1_col0\" class=\"data row1 col0\" >1957-12</td>\n",
       "                        <td id=\"T_f54a4_row1_col1\" class=\"data row1 col1\" >47.2819</td>\n",
       "                        <td id=\"T_f54a4_row1_col2\" class=\"data row1 col2\" >4743.9295</td>\n",
       "                        <td id=\"T_f54a4_row1_col3\" class=\"data row1 col3\" >0.1108</td>\n",
       "                        <td id=\"T_f54a4_row1_col4\" class=\"data row1 col4\" >0.1205</td>\n",
       "                        <td id=\"T_f54a4_row1_col5\" class=\"data row1 col5\" >-0.2428</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_f54a4_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_f54a4_row2_col0\" class=\"data row2 col0\" >1958-12</td>\n",
       "                        <td id=\"T_f54a4_row2_col1\" class=\"data row2 col1\" >79.0434</td>\n",
       "                        <td id=\"T_f54a4_row2_col2\" class=\"data row2 col2\" >10522.4321</td>\n",
       "                        <td id=\"T_f54a4_row2_col3\" class=\"data row2 col3\" >0.1667</td>\n",
       "                        <td id=\"T_f54a4_row2_col4\" class=\"data row2 col4\" >0.1909</td>\n",
       "                        <td id=\"T_f54a4_row2_col5\" class=\"data row2 col5\" >-1.3541</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_f54a4_level0_row3\" class=\"row_heading level0 row3\" >Mean</th>\n",
       "                        <td id=\"T_f54a4_row3_col0\" class=\"data row3 col0\" >nan</td>\n",
       "                        <td id=\"T_f54a4_row3_col1\" class=\"data row3 col1\" >60.2976</td>\n",
       "                        <td id=\"T_f54a4_row3_col2\" class=\"data row3 col2\" >6885.4347</td>\n",
       "                        <td id=\"T_f54a4_row3_col3\" class=\"data row3 col3\" >0.1369</td>\n",
       "                        <td id=\"T_f54a4_row3_col4\" class=\"data row3 col4\" >0.1534</td>\n",
       "                        <td id=\"T_f54a4_row3_col5\" class=\"data row3 col5\" >-0.7838</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_f54a4_level0_row4\" class=\"row_heading level0 row4\" >SD</th>\n",
       "                        <td id=\"T_f54a4_row4_col0\" class=\"data row4 col0\" >nan</td>\n",
       "                        <td id=\"T_f54a4_row4_col1\" class=\"data row4 col1\" >13.5849</td>\n",
       "                        <td id=\"T_f54a4_row4_col2\" class=\"data row4 col2\" >2585.2332</td>\n",
       "                        <td id=\"T_f54a4_row4_col3\" class=\"data row4 col3\" >0.0230</td>\n",
       "                        <td id=\"T_f54a4_row4_col4\" class=\"data row4 col4\" >0.0289</td>\n",
       "                        <td id=\"T_f54a4_row4_col5\" class=\"data row4 col5\" >0.4542</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e580e02f08>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "NaiveForecaster(sp=1, strategy='drift', window_length=None)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = exp.create_model(\"naive\", strategy='drift')\n",
    "model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Change number of folds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_75e1e_row5_col0,#T_75e1e_row5_col1,#T_75e1e_row5_col2,#T_75e1e_row5_col3,#T_75e1e_row5_col4,#T_75e1e_row5_col5{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_75e1e_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >cutoff</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_75e1e_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_75e1e_row0_col0\" class=\"data row0 col0\" >1954-12</td>\n",
       "                        <td id=\"T_75e1e_row0_col1\" class=\"data row0 col1\" >55.0000</td>\n",
       "                        <td id=\"T_75e1e_row0_col2\" class=\"data row0 col2\" >4652.8333</td>\n",
       "                        <td id=\"T_75e1e_row0_col3\" class=\"data row0 col3\" >0.1782</td>\n",
       "                        <td id=\"T_75e1e_row0_col4\" class=\"data row0 col4\" >0.2038</td>\n",
       "                        <td id=\"T_75e1e_row0_col5\" class=\"data row0 col5\" >-1.8583</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_75e1e_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_75e1e_row1_col0\" class=\"data row1 col0\" >1955-12</td>\n",
       "                        <td id=\"T_75e1e_row1_col1\" class=\"data row1 col1\" >51.5833</td>\n",
       "                        <td id=\"T_75e1e_row1_col2\" class=\"data row1 col2\" >4624.9167</td>\n",
       "                        <td id=\"T_75e1e_row1_col3\" class=\"data row1 col3\" >0.1424</td>\n",
       "                        <td id=\"T_75e1e_row1_col4\" class=\"data row1 col4\" >0.1605</td>\n",
       "                        <td id=\"T_75e1e_row1_col5\" class=\"data row1 col5\" >-1.2025</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_75e1e_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_75e1e_row2_col0\" class=\"data row2 col0\" >1956-12</td>\n",
       "                        <td id=\"T_75e1e_row2_col1\" class=\"data row2 col1\" >63.4167</td>\n",
       "                        <td id=\"T_75e1e_row2_col2\" class=\"data row2 col2\" >6967.9167</td>\n",
       "                        <td id=\"T_75e1e_row2_col3\" class=\"data row2 col3\" >0.1549</td>\n",
       "                        <td id=\"T_75e1e_row2_col4\" class=\"data row2 col4\" >0.1766</td>\n",
       "                        <td id=\"T_75e1e_row2_col5\" class=\"data row2 col5\" >-1.2681</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_75e1e_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_75e1e_row3_col0\" class=\"data row3 col0\" >1957-12</td>\n",
       "                        <td id=\"T_75e1e_row3_col1\" class=\"data row3 col1\" >52.3333</td>\n",
       "                        <td id=\"T_75e1e_row3_col2\" class=\"data row3 col2\" >5842.1667</td>\n",
       "                        <td id=\"T_75e1e_row3_col3\" class=\"data row3 col3\" >0.1205</td>\n",
       "                        <td id=\"T_75e1e_row3_col4\" class=\"data row3 col4\" >0.1351</td>\n",
       "                        <td id=\"T_75e1e_row3_col5\" class=\"data row3 col5\" >-0.5305</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_75e1e_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_75e1e_row4_col0\" class=\"data row4 col0\" >1958-12</td>\n",
       "                        <td id=\"T_75e1e_row4_col1\" class=\"data row4 col1\" >91.3333</td>\n",
       "                        <td id=\"T_75e1e_row4_col2\" class=\"data row4 col2\" >12811.6667</td>\n",
       "                        <td id=\"T_75e1e_row4_col3\" class=\"data row4 col3\" >0.1954</td>\n",
       "                        <td id=\"T_75e1e_row4_col4\" class=\"data row4 col4\" >0.2259</td>\n",
       "                        <td id=\"T_75e1e_row4_col5\" class=\"data row4 col5\" >-1.8662</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_75e1e_level0_row5\" class=\"row_heading level0 row5\" >Mean</th>\n",
       "                        <td id=\"T_75e1e_row5_col0\" class=\"data row5 col0\" >nan</td>\n",
       "                        <td id=\"T_75e1e_row5_col1\" class=\"data row5 col1\" >62.7333</td>\n",
       "                        <td id=\"T_75e1e_row5_col2\" class=\"data row5 col2\" >6979.9000</td>\n",
       "                        <td id=\"T_75e1e_row5_col3\" class=\"data row5 col3\" >0.1583</td>\n",
       "                        <td id=\"T_75e1e_row5_col4\" class=\"data row5 col4\" >0.1804</td>\n",
       "                        <td id=\"T_75e1e_row5_col5\" class=\"data row5 col5\" >-1.3451</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_75e1e_level0_row6\" class=\"row_heading level0 row6\" >SD</th>\n",
       "                        <td id=\"T_75e1e_row6_col0\" class=\"data row6 col0\" >nan</td>\n",
       "                        <td id=\"T_75e1e_row6_col1\" class=\"data row6 col1\" >14.9045</td>\n",
       "                        <td id=\"T_75e1e_row6_col2\" class=\"data row6 col2\" >3041.8833</td>\n",
       "                        <td id=\"T_75e1e_row6_col3\" class=\"data row6 col3\" >0.0263</td>\n",
       "                        <td id=\"T_75e1e_row6_col4\" class=\"data row6 col4\" >0.0319</td>\n",
       "                        <td id=\"T_75e1e_row6_col5\" class=\"data row6 col5\" >0.4949</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e58009de48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "NaiveForecaster(sp=1, strategy='last', window_length=None)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = exp.create_model(\"naive\", fold=5)\n",
    "model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Tune Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Returning the Tuner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_757b3_row3_col0,#T_757b3_row3_col1,#T_757b3_row3_col2,#T_757b3_row3_col3,#T_757b3_row3_col4,#T_757b3_row3_col5{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_757b3_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >cutoff</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_757b3_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_757b3_row0_col0\" class=\"data row0 col0\" >1956-12</td>\n",
       "                        <td id=\"T_757b3_row0_col1\" class=\"data row0 col1\" >54.5675</td>\n",
       "                        <td id=\"T_757b3_row0_col2\" class=\"data row0 col2\" >5389.9425</td>\n",
       "                        <td id=\"T_757b3_row0_col3\" class=\"data row0 col3\" >0.1332</td>\n",
       "                        <td id=\"T_757b3_row0_col4\" class=\"data row0 col4\" >0.1489</td>\n",
       "                        <td id=\"T_757b3_row0_col5\" class=\"data row0 col5\" >-0.7545</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_757b3_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_757b3_row1_col0\" class=\"data row1 col0\" >1957-12</td>\n",
       "                        <td id=\"T_757b3_row1_col1\" class=\"data row1 col1\" >47.2819</td>\n",
       "                        <td id=\"T_757b3_row1_col2\" class=\"data row1 col2\" >4743.9295</td>\n",
       "                        <td id=\"T_757b3_row1_col3\" class=\"data row1 col3\" >0.1108</td>\n",
       "                        <td id=\"T_757b3_row1_col4\" class=\"data row1 col4\" >0.1205</td>\n",
       "                        <td id=\"T_757b3_row1_col5\" class=\"data row1 col5\" >-0.2428</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_757b3_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_757b3_row2_col0\" class=\"data row2 col0\" >1958-12</td>\n",
       "                        <td id=\"T_757b3_row2_col1\" class=\"data row2 col1\" >79.0434</td>\n",
       "                        <td id=\"T_757b3_row2_col2\" class=\"data row2 col2\" >10522.4321</td>\n",
       "                        <td id=\"T_757b3_row2_col3\" class=\"data row2 col3\" >0.1667</td>\n",
       "                        <td id=\"T_757b3_row2_col4\" class=\"data row2 col4\" >0.1909</td>\n",
       "                        <td id=\"T_757b3_row2_col5\" class=\"data row2 col5\" >-1.3541</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_757b3_level0_row3\" class=\"row_heading level0 row3\" >Mean</th>\n",
       "                        <td id=\"T_757b3_row3_col0\" class=\"data row3 col0\" >nan</td>\n",
       "                        <td id=\"T_757b3_row3_col1\" class=\"data row3 col1\" >60.2976</td>\n",
       "                        <td id=\"T_757b3_row3_col2\" class=\"data row3 col2\" >6885.4347</td>\n",
       "                        <td id=\"T_757b3_row3_col3\" class=\"data row3 col3\" >0.1369</td>\n",
       "                        <td id=\"T_757b3_row3_col4\" class=\"data row3 col4\" >0.1534</td>\n",
       "                        <td id=\"T_757b3_row3_col5\" class=\"data row3 col5\" >-0.7838</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_757b3_level0_row4\" class=\"row_heading level0 row4\" >SD</th>\n",
       "                        <td id=\"T_757b3_row4_col0\" class=\"data row4 col0\" >nan</td>\n",
       "                        <td id=\"T_757b3_row4_col1\" class=\"data row4 col1\" >13.5849</td>\n",
       "                        <td id=\"T_757b3_row4_col2\" class=\"data row4 col2\" >2585.2332</td>\n",
       "                        <td id=\"T_757b3_row4_col3\" class=\"data row4 col3\" >0.0230</td>\n",
       "                        <td id=\"T_757b3_row4_col4\" class=\"data row4 col4\" >0.0289</td>\n",
       "                        <td id=\"T_757b3_row4_col5\" class=\"data row4 col5\" >0.4542</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e5a93e4988>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tuned_model, tuner = exp.tune_model(model, return_tuner=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check the results of the tuner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_window_length</th>\n",
       "      <th>param_strategy</th>\n",
       "      <th>param_sp</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_smape</th>\n",
       "      <th>split1_test_smape</th>\n",
       "      <th>split2_test_smape</th>\n",
       "      <th>mean_test_smape</th>\n",
       "      <th>std_test_smape</th>\n",
       "      <th>rank_test_smape</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.000332</td>\n",
       "      <td>0.000469</td>\n",
       "      <td>0.001995</td>\n",
       "      <td>6.257699e-07</td>\n",
       "      <td>None</td>\n",
       "      <td>last</td>\n",
       "      <td>1</td>\n",
       "      <td>{'window_length': None, 'strategy': 'last', 's...</td>\n",
       "      <td>0.176606</td>\n",
       "      <td>0.135060</td>\n",
       "      <td>0.225903</td>\n",
       "      <td>0.179190</td>\n",
       "      <td>0.037132</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.000666</td>\n",
       "      <td>0.000471</td>\n",
       "      <td>0.001653</td>\n",
       "      <td>9.473634e-04</td>\n",
       "      <td>None</td>\n",
       "      <td>mean</td>\n",
       "      <td>1</td>\n",
       "      <td>{'window_length': None, 'strategy': 'mean', 's...</td>\n",
       "      <td>0.518785</td>\n",
       "      <td>0.476298</td>\n",
       "      <td>0.527557</td>\n",
       "      <td>0.507547</td>\n",
       "      <td>0.022385</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.001652</td>\n",
       "      <td>4.822561e-04</td>\n",
       "      <td>None</td>\n",
       "      <td>drift</td>\n",
       "      <td>1</td>\n",
       "      <td>{'window_length': None, 'strategy': 'drift', '...</td>\n",
       "      <td>0.148905</td>\n",
       "      <td>0.120522</td>\n",
       "      <td>0.190900</td>\n",
       "      <td>0.153442</td>\n",
       "      <td>0.028910</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "0       0.000332      0.000469         0.001995    6.257699e-07   \n",
       "1       0.000666      0.000471         0.001653    9.473634e-04   \n",
       "2       0.000000      0.000000         0.001652    4.822561e-04   \n",
       "\n",
       "  param_window_length param_strategy param_sp  \\\n",
       "0                None           last        1   \n",
       "1                None           mean        1   \n",
       "2                None          drift        1   \n",
       "\n",
       "                                              params  split0_test_smape  \\\n",
       "0  {'window_length': None, 'strategy': 'last', 's...           0.176606   \n",
       "1  {'window_length': None, 'strategy': 'mean', 's...           0.518785   \n",
       "2  {'window_length': None, 'strategy': 'drift', '...           0.148905   \n",
       "\n",
       "   split1_test_smape  split2_test_smape  mean_test_smape  std_test_smape  \\\n",
       "0           0.135060           0.225903         0.179190        0.037132   \n",
       "1           0.476298           0.527557         0.507547        0.022385   \n",
       "2           0.120522           0.190900         0.153442        0.028910   \n",
       "\n",
       "   rank_test_smape  \n",
       "0                2  \n",
       "1                3  \n",
       "2                1  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grid Results\n",
    "pd.DataFrame(tuner.cv_results_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'sp': 1, 'strategy': 'drift', 'window_length': None}\n"
     ]
    }
   ],
   "source": [
    "# Best Hyperparameters\n",
    "pprint(tuner.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Defining a Custom Grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'sp': [12],\n",
       " 'strategy': ['last', 'mean', 'drift'],\n",
       " 'window_length': [12, 24, None]}"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my_grid = {\n",
    "    'sp': [12],\n",
    "    'strategy': ['last', 'mean', 'drift'],\n",
    "    'window_length': [12, 24, None]\n",
    "}\n",
    "my_grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_c4fe5_row3_col0,#T_c4fe5_row3_col1,#T_c4fe5_row3_col2,#T_c4fe5_row3_col3,#T_c4fe5_row3_col4,#T_c4fe5_row3_col5{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_c4fe5_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >cutoff</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_c4fe5_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_c4fe5_row0_col0\" class=\"data row0 col0\" >1956-12</td>\n",
       "                        <td id=\"T_c4fe5_row0_col1\" class=\"data row0 col1\" >40.1667</td>\n",
       "                        <td id=\"T_c4fe5_row0_col2\" class=\"data row0 col2\" >1720.1667</td>\n",
       "                        <td id=\"T_c4fe5_row0_col3\" class=\"data row0 col3\" >0.1076</td>\n",
       "                        <td id=\"T_c4fe5_row0_col4\" class=\"data row0 col4\" >0.1138</td>\n",
       "                        <td id=\"T_c4fe5_row0_col5\" class=\"data row0 col5\" >0.4401</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_c4fe5_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_c4fe5_row1_col0\" class=\"data row1 col0\" >1957-12</td>\n",
       "                        <td id=\"T_c4fe5_row1_col1\" class=\"data row1 col1\" >12.5833</td>\n",
       "                        <td id=\"T_c4fe5_row1_col2\" class=\"data row1 col2\" >289.4167</td>\n",
       "                        <td id=\"T_c4fe5_row1_col3\" class=\"data row1 col3\" >0.0314</td>\n",
       "                        <td id=\"T_c4fe5_row1_col4\" class=\"data row1 col4\" >0.0322</td>\n",
       "                        <td id=\"T_c4fe5_row1_col5\" class=\"data row1 col5\" >0.9242</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_c4fe5_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_c4fe5_row2_col0\" class=\"data row2 col0\" >1958-12</td>\n",
       "                        <td id=\"T_c4fe5_row2_col1\" class=\"data row2 col1\" >47.3333</td>\n",
       "                        <td id=\"T_c4fe5_row2_col2\" class=\"data row2 col2\" >2426.0000</td>\n",
       "                        <td id=\"T_c4fe5_row2_col3\" class=\"data row2 col3\" >0.1106</td>\n",
       "                        <td id=\"T_c4fe5_row2_col4\" class=\"data row2 col4\" >0.1176</td>\n",
       "                        <td id=\"T_c4fe5_row2_col5\" class=\"data row2 col5\" >0.4573</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_c4fe5_level0_row3\" class=\"row_heading level0 row3\" >Mean</th>\n",
       "                        <td id=\"T_c4fe5_row3_col0\" class=\"data row3 col0\" >nan</td>\n",
       "                        <td id=\"T_c4fe5_row3_col1\" class=\"data row3 col1\" >33.3611</td>\n",
       "                        <td id=\"T_c4fe5_row3_col2\" class=\"data row3 col2\" >1478.5278</td>\n",
       "                        <td id=\"T_c4fe5_row3_col3\" class=\"data row3 col3\" >0.0832</td>\n",
       "                        <td id=\"T_c4fe5_row3_col4\" class=\"data row3 col4\" >0.0879</td>\n",
       "                        <td id=\"T_c4fe5_row3_col5\" class=\"data row3 col5\" >0.6072</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_c4fe5_level0_row4\" class=\"row_heading level0 row4\" >SD</th>\n",
       "                        <td id=\"T_c4fe5_row4_col0\" class=\"data row4 col0\" >nan</td>\n",
       "                        <td id=\"T_c4fe5_row4_col1\" class=\"data row4 col1\" >14.9806</td>\n",
       "                        <td id=\"T_c4fe5_row4_col2\" class=\"data row4 col2\" >888.8341</td>\n",
       "                        <td id=\"T_c4fe5_row4_col3\" class=\"data row4 col3\" >0.0367</td>\n",
       "                        <td id=\"T_c4fe5_row4_col4\" class=\"data row4 col4\" >0.0394</td>\n",
       "                        <td id=\"T_c4fe5_row4_col5\" class=\"data row4 col5\" >0.2243</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e5a93d4e48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tuned_model = exp.tune_model(model, custom_grid=my_grid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The performance improved over the original tuning (we essentially converted the naive model to a seasonal naive model with this new grid)**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prediction Intervals\n",
    "\n",
    "Available for some models only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_09894_row3_col0,#T_09894_row3_col1,#T_09894_row3_col2,#T_09894_row3_col3,#T_09894_row3_col4,#T_09894_row3_col5{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_09894_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >cutoff</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_09894_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_09894_row0_col0\" class=\"data row0 col0\" >1956-12</td>\n",
       "                        <td id=\"T_09894_row0_col1\" class=\"data row0 col1\" >57.5799</td>\n",
       "                        <td id=\"T_09894_row0_col2\" class=\"data row0 col2\" >5883.9192</td>\n",
       "                        <td id=\"T_09894_row0_col3\" class=\"data row0 col3\" >0.1406</td>\n",
       "                        <td id=\"T_09894_row0_col4\" class=\"data row0 col4\" >0.1582</td>\n",
       "                        <td id=\"T_09894_row0_col5\" class=\"data row0 col5\" >-0.9153</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_09894_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_09894_row1_col0\" class=\"data row1 col0\" >1957-12</td>\n",
       "                        <td id=\"T_09894_row1_col1\" class=\"data row1 col1\" >48.6352</td>\n",
       "                        <td id=\"T_09894_row1_col2\" class=\"data row1 col2\" >5040.6939</td>\n",
       "                        <td id=\"T_09894_row1_col3\" class=\"data row1 col3\" >0.1131</td>\n",
       "                        <td id=\"T_09894_row1_col4\" class=\"data row1 col4\" >0.1244</td>\n",
       "                        <td id=\"T_09894_row1_col5\" class=\"data row1 col5\" >-0.3205</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_09894_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_09894_row2_col0\" class=\"data row2 col0\" >1958-12</td>\n",
       "                        <td id=\"T_09894_row2_col1\" class=\"data row2 col1\" >82.1049</td>\n",
       "                        <td id=\"T_09894_row2_col2\" class=\"data row2 col2\" >11072.6778</td>\n",
       "                        <td id=\"T_09894_row2_col3\" class=\"data row2 col3\" >0.1738</td>\n",
       "                        <td id=\"T_09894_row2_col4\" class=\"data row2 col4\" >0.1995</td>\n",
       "                        <td id=\"T_09894_row2_col5\" class=\"data row2 col5\" >-1.4772</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_09894_level0_row3\" class=\"row_heading level0 row3\" >Mean</th>\n",
       "                        <td id=\"T_09894_row3_col0\" class=\"data row3 col0\" >nan</td>\n",
       "                        <td id=\"T_09894_row3_col1\" class=\"data row3 col1\" >62.7733</td>\n",
       "                        <td id=\"T_09894_row3_col2\" class=\"data row3 col2\" >7332.4303</td>\n",
       "                        <td id=\"T_09894_row3_col3\" class=\"data row3 col3\" >0.1425</td>\n",
       "                        <td id=\"T_09894_row3_col4\" class=\"data row3 col4\" >0.1607</td>\n",
       "                        <td id=\"T_09894_row3_col5\" class=\"data row3 col5\" >-0.9043</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_09894_level0_row4\" class=\"row_heading level0 row4\" >SD</th>\n",
       "                        <td id=\"T_09894_row4_col0\" class=\"data row4 col0\" >nan</td>\n",
       "                        <td id=\"T_09894_row4_col1\" class=\"data row4 col1\" >14.1488</td>\n",
       "                        <td id=\"T_09894_row4_col2\" class=\"data row4 col2\" >2667.0640</td>\n",
       "                        <td id=\"T_09894_row4_col3\" class=\"data row4 col3\" >0.0248</td>\n",
       "                        <td id=\"T_09894_row4_col4\" class=\"data row4 col4\" >0.0307</td>\n",
       "                        <td id=\"T_09894_row4_col5\" class=\"data row4 col5\" >0.4723</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e580be0948>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = exp.create_model(\"theta\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "</style><table id=\"T_4e6e9_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Model</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_4e6e9_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_4e6e9_row0_col0\" class=\"data row0 col0\" >Theta Forecaster</td>\n",
       "                        <td id=\"T_4e6e9_row0_col1\" class=\"data row0 col1\" >69.7290</td>\n",
       "                        <td id=\"T_4e6e9_row0_col2\" class=\"data row0 col2\" >96.1663</td>\n",
       "                        <td id=\"T_4e6e9_row0_col3\" class=\"data row0 col3\" >0.1306</td>\n",
       "                        <td id=\"T_4e6e9_row0_col4\" class=\"data row0 col4\" >0.1462</td>\n",
       "                        <td id=\"T_4e6e9_row0_col5\" class=\"data row0 col5\" >-0.6695</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e580b2e3c8>"
      ]
     },
     "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>y_pred</th>\n",
       "      <th>lower</th>\n",
       "      <th>upper</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1960-01</th>\n",
       "      <td>407.3563</td>\n",
       "      <td>320.5866</td>\n",
       "      <td>494.1260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-02</th>\n",
       "      <td>408.6381</td>\n",
       "      <td>302.4562</td>\n",
       "      <td>514.8201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-03</th>\n",
       "      <td>409.9200</td>\n",
       "      <td>287.3630</td>\n",
       "      <td>532.4770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-04</th>\n",
       "      <td>411.2019</td>\n",
       "      <td>274.2133</td>\n",
       "      <td>548.1904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-05</th>\n",
       "      <td>412.4837</td>\n",
       "      <td>262.4455</td>\n",
       "      <td>562.5220</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-06</th>\n",
       "      <td>413.7656</td>\n",
       "      <td>251.7252</td>\n",
       "      <td>575.8060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-07</th>\n",
       "      <td>415.0474</td>\n",
       "      <td>241.8345</td>\n",
       "      <td>588.2604</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-08</th>\n",
       "      <td>416.3293</td>\n",
       "      <td>232.6221</td>\n",
       "      <td>600.0365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-09</th>\n",
       "      <td>417.6111</td>\n",
       "      <td>223.9776</td>\n",
       "      <td>611.2447</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-10</th>\n",
       "      <td>418.8930</td>\n",
       "      <td>215.8177</td>\n",
       "      <td>621.9683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-11</th>\n",
       "      <td>420.1749</td>\n",
       "      <td>208.0777</td>\n",
       "      <td>632.2720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-12</th>\n",
       "      <td>421.4567</td>\n",
       "      <td>200.7061</td>\n",
       "      <td>642.2073</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           y_pred     lower     upper\n",
       "1960-01  407.3563  320.5866  494.1260\n",
       "1960-02  408.6381  302.4562  514.8201\n",
       "1960-03  409.9200  287.3630  532.4770\n",
       "1960-04  411.2019  274.2133  548.1904\n",
       "1960-05  412.4837  262.4455  562.5220\n",
       "1960-06  413.7656  251.7252  575.8060\n",
       "1960-07  415.0474  241.8345  588.2604\n",
       "1960-08  416.3293  232.6221  600.0365\n",
       "1960-09  417.6111  223.9776  611.2447\n",
       "1960-10  418.8930  215.8177  621.9683\n",
       "1960-11  420.1749  208.0777  632.2720\n",
       "1960-12  421.4567  200.7061  642.2073"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exp.predict_model(model, return_pred_int=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Models that do not provide a predicton interval simply return NA values**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "#T_7f385_row3_col0,#T_7f385_row3_col1,#T_7f385_row3_col2,#T_7f385_row3_col3,#T_7f385_row3_col4,#T_7f385_row3_col5{\n",
       "            background:  yellow;\n",
       "        }</style><table id=\"T_7f385_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >cutoff</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_7f385_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_7f385_row0_col0\" class=\"data row0 col0\" >1956-12</td>\n",
       "                        <td id=\"T_7f385_row0_col1\" class=\"data row0 col1\" >63.4167</td>\n",
       "                        <td id=\"T_7f385_row0_col2\" class=\"data row0 col2\" >6967.9167</td>\n",
       "                        <td id=\"T_7f385_row0_col3\" class=\"data row0 col3\" >0.1549</td>\n",
       "                        <td id=\"T_7f385_row0_col4\" class=\"data row0 col4\" >0.1766</td>\n",
       "                        <td id=\"T_7f385_row0_col5\" class=\"data row0 col5\" >-1.2681</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7f385_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_7f385_row1_col0\" class=\"data row1 col0\" >1957-12</td>\n",
       "                        <td id=\"T_7f385_row1_col1\" class=\"data row1 col1\" >52.3333</td>\n",
       "                        <td id=\"T_7f385_row1_col2\" class=\"data row1 col2\" >5842.1667</td>\n",
       "                        <td id=\"T_7f385_row1_col3\" class=\"data row1 col3\" >0.1205</td>\n",
       "                        <td id=\"T_7f385_row1_col4\" class=\"data row1 col4\" >0.1351</td>\n",
       "                        <td id=\"T_7f385_row1_col5\" class=\"data row1 col5\" >-0.5305</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7f385_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_7f385_row2_col0\" class=\"data row2 col0\" >1958-12</td>\n",
       "                        <td id=\"T_7f385_row2_col1\" class=\"data row2 col1\" >91.3333</td>\n",
       "                        <td id=\"T_7f385_row2_col2\" class=\"data row2 col2\" >12811.6667</td>\n",
       "                        <td id=\"T_7f385_row2_col3\" class=\"data row2 col3\" >0.1954</td>\n",
       "                        <td id=\"T_7f385_row2_col4\" class=\"data row2 col4\" >0.2259</td>\n",
       "                        <td id=\"T_7f385_row2_col5\" class=\"data row2 col5\" >-1.8662</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7f385_level0_row3\" class=\"row_heading level0 row3\" >Mean</th>\n",
       "                        <td id=\"T_7f385_row3_col0\" class=\"data row3 col0\" >nan</td>\n",
       "                        <td id=\"T_7f385_row3_col1\" class=\"data row3 col1\" >69.0278</td>\n",
       "                        <td id=\"T_7f385_row3_col2\" class=\"data row3 col2\" >8540.5833</td>\n",
       "                        <td id=\"T_7f385_row3_col3\" class=\"data row3 col3\" >0.1569</td>\n",
       "                        <td id=\"T_7f385_row3_col4\" class=\"data row3 col4\" >0.1792</td>\n",
       "                        <td id=\"T_7f385_row3_col5\" class=\"data row3 col5\" >-1.2216</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_7f385_level0_row4\" class=\"row_heading level0 row4\" >SD</th>\n",
       "                        <td id=\"T_7f385_row4_col0\" class=\"data row4 col0\" >nan</td>\n",
       "                        <td id=\"T_7f385_row4_col1\" class=\"data row4 col1\" >16.4086</td>\n",
       "                        <td id=\"T_7f385_row4_col2\" class=\"data row4 col2\" >3054.8806</td>\n",
       "                        <td id=\"T_7f385_row4_col3\" class=\"data row4 col3\" >0.0306</td>\n",
       "                        <td id=\"T_7f385_row4_col4\" class=\"data row4 col4\" >0.0371</td>\n",
       "                        <td id=\"T_7f385_row4_col5\" class=\"data row4 col5\" >0.5463</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e580a34b48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = exp.create_model(\"naive\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "</style><table id=\"T_4080f_\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Model</th>        <th class=\"col_heading level0 col1\" >MAE</th>        <th class=\"col_heading level0 col2\" >RMSE</th>        <th class=\"col_heading level0 col3\" >MAPE</th>        <th class=\"col_heading level0 col4\" >SMAPE</th>        <th class=\"col_heading level0 col5\" >R2</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_4080f_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_4080f_row0_col0\" class=\"data row0 col0\" >Naive Forecaster</td>\n",
       "                        <td id=\"T_4080f_row0_col1\" class=\"data row0 col1\" >76.0000</td>\n",
       "                        <td id=\"T_4080f_row0_col2\" class=\"data row0 col2\" >102.9765</td>\n",
       "                        <td id=\"T_4080f_row0_col3\" class=\"data row0 col3\" >0.1425</td>\n",
       "                        <td id=\"T_4080f_row0_col4\" class=\"data row0 col4\" >0.1612</td>\n",
       "                        <td id=\"T_4080f_row0_col5\" class=\"data row0 col5\" >-0.9143</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1e580c23f88>"
      ]
     },
     "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>y_pred</th>\n",
       "      <th>lower</th>\n",
       "      <th>upper</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1960-01</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-02</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-03</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-04</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-05</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-06</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-07</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-08</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-09</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-10</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-11</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-12</th>\n",
       "      <td>405.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         y_pred  lower  upper\n",
       "1960-01   405.0    NaN    NaN\n",
       "1960-02   405.0    NaN    NaN\n",
       "1960-03   405.0    NaN    NaN\n",
       "1960-04   405.0    NaN    NaN\n",
       "1960-05   405.0    NaN    NaN\n",
       "1960-06   405.0    NaN    NaN\n",
       "1960-07   405.0    NaN    NaN\n",
       "1960-08   405.0    NaN    NaN\n",
       "1960-09   405.0    NaN    NaN\n",
       "1960-10   405.0    NaN    NaN\n",
       "1960-11   405.0    NaN    NaN\n",
       "1960-12   405.0    NaN    NaN"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exp.predict_model(model, return_pred_int=True)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "2e79732785b968a887869d670772af3eb35a449ef73721b650f9c2f4a3fd4bbe"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}