ChadFulton/automatic.ipynb

## automatic.ipynb
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# statsmodels.tsa.automatic\n",
    "\n",
    "The `automatic` package in Statsmodels' time series package helps to quickly select a reasonable model for a given time series that can be used for analysis or forecasting. It consists of three components:\n",
    "\n",
    "- One-step model selection\n",
    "- Specification search\n",
    "- \"Time series cross-validation\" (in-development, not described here yet)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import statsmodels.api as sm\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "dta = sm.datasets.macrodata.load_pandas().data\n",
    "dta.index = pd.PeriodIndex(start='1959Q1', end='2009Q3', freq='Q')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## One-step model selection\n",
    "\n",
    "Automatic model selection performs the specification search, selects the best model, and returns the fitted results from that model, all in one step. It is available for exponential smoothing models, SARIMAX models, and unobserved components models.\n",
    "\n",
    "It does not provide the user with any feedback on the selection search itself (see below for informaton on the `selection_search` functions that provide detailed information about the search). Instead, it attempts to provide the user immediately with a \"good\" model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                       ExponentialSmoothing Model Results                       \n",
      "================================================================================\n",
      "Dep. Variable:                    endog   No. Observations:                  203\n",
      "Model:             ExponentialSmoothing   SSE                            131.169\n",
      "Optimized:                         True   AIC                            -80.653\n",
      "Trend:                         Additive   BIC                            -67.401\n",
      "Seasonal:                          None   AICC                           -80.225\n",
      "Seasonal Periods:                  None   Date:                 Sat, 08 Sep 2018\n",
      "Box-Cox:                          False   Time:                         23:37:50\n",
      "Box-Cox Coeff.:                    None                                         \n",
      "==============================================================================\n",
      "                       coeff                 code              optimized      \n",
      "------------------------------------------------------------------------------\n",
      "smoothing_level            1.0000000                alpha                 True\n",
      "smoothing_slope            0.1424878                 beta                 True\n",
      "initial_level              28.864420                  l.0                 True\n",
      "initial_slope              0.1155756                  b.0                 True\n",
      "------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# Exponential smoothing\n",
    "res_es = sm.tsa.automatic.auto_es(dta.cpi)\n",
    "print(res_es.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                           Statespace Model Results                           \n",
      "==============================================================================\n",
      "Dep. Variable:                    cpi   No. Observations:                  203\n",
      "Model:               SARIMAX(3, 2, 3)   Log Likelihood                -230.690\n",
      "Date:                Sat, 08 Sep 2018   AIC                            475.379\n",
      "Time:                        23:37:54   BIC                            498.502\n",
      "Sample:                    03-31-1959   HQIC                           484.736\n",
      "                         - 09-30-2009                                         \n",
      "Covariance Type:                  opg                                         \n",
      "==============================================================================\n",
      "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "ar.L1         -1.1686      0.056    -20.714      0.000      -1.279      -1.058\n",
      "ar.L2         -0.6785      0.099     -6.840      0.000      -0.873      -0.484\n",
      "ar.L3          0.1899      0.083      2.300      0.021       0.028       0.352\n",
      "ma.L1          0.5499      0.036     15.424      0.000       0.480       0.620\n",
      "ma.L2         -0.3612      0.040     -9.145      0.000      -0.439      -0.284\n",
      "ma.L3         -0.8869      0.041    -21.748      0.000      -0.967      -0.807\n",
      "sigma2         0.5702      0.042     13.573      0.000       0.488       0.653\n",
      "===================================================================================\n",
      "Ljung-Box (Q):                       20.79   Jarque-Bera (JB):              1515.00\n",
      "Prob(Q):                              0.99   Prob(JB):                         0.00\n",
      "Heteroskedasticity (H):              19.11   Skew:                            -1.82\n",
      "Prob(H) (two-sided):                  0.00   Kurtosis:                        15.95\n",
      "===================================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Covariance matrix calculated using the outer product of gradients (complex-step).\n"
     ]
    }
   ],
   "source": [
    "# SARIMAX\n",
    "res_sarimax = sm.tsa.automatic.auto_sarimax(dta.cpi, d=2, stepwise=True, fit_kwargs={'maxiter': 100})\n",
    "print(res_sarimax.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                        Unobserved Components Results                         \n",
      "==============================================================================\n",
      "Dep. Variable:                    cpi   No. Observations:                  203\n",
      "Model:             local linear trend   Log Likelihood                -242.964\n",
      "Date:                Sat, 08 Sep 2018   AIC                            491.928\n",
      "Time:                        23:37:54   BIC                            501.838\n",
      "Sample:                    03-31-1959   HQIC                           495.938\n",
      "                         - 09-30-2009                                         \n",
      "Covariance Type:                  opg                                         \n",
      "====================================================================================\n",
      "                       coef    std err          z      P>|z|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------------\n",
      "sigma2.irregular  9.238e-11      0.021   4.43e-09      1.000      -0.041       0.041\n",
      "sigma2.level         0.5546      0.043     12.940      0.000       0.471       0.639\n",
      "sigma2.trend         0.0147      0.007      2.247      0.025       0.002       0.028\n",
      "===================================================================================\n",
      "Ljung-Box (Q):                       30.71   Jarque-Bera (JB):              4231.58\n",
      "Prob(Q):                              0.85   Prob(JB):                         0.00\n",
      "Heteroskedasticity (H):              23.57   Skew:                            -2.61\n",
      "Prob(H) (two-sided):                  0.00   Kurtosis:                        24.86\n",
      "===================================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Covariance matrix calculated using the outer product of gradients (complex-step).\n"
     ]
    }
   ],
   "source": [
    "# Unobserved components\n",
    "res_uc = sm.tsa.automatic.auto_uc(dta.cpi, level='I(2)')\n",
    "print(res_uc.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compare results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean squared error (in-sample):\n",
      "Exponential smoothing: 0.6526\n",
      "SARIMAX:               0.5710\n",
      "Unobserved components: 0.6527\n"
     ]
    }
   ],
   "source": [
    "print('Mean squared error (in-sample):')\n",
    "print('Exponential smoothing: %.4f' % np.mean(res_es.resid[2:]**2))\n",
    "print('SARIMAX:               %.4f' % np.mean(res_sarimax.resid[2:]**2))\n",
    "print('Unobserved components: %.4f' % np.mean(res_uc.resid[2:]**2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(13, 5))\n",
    "\n",
    "dta.cpi.loc['2006':].plot(ax=ax, label='Data')\n",
    "res_es.forecast(10).plot(ax=ax, label='Exponential smoothing')\n",
    "res_sarimax.forecast(10).plot(ax=ax, label='SARIMAX')\n",
    "res_uc.forecast(10).plot(ax=ax, label='Unobserved components')\n",
    "\n",
    "ax.set(title='Forecasts from automatically selected models')\n",
    "ax.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Alternative measure\n",
    "\n",
    "The default selection measure is the Akaike information criteria (`measure='aic'`), but once can also use the Bayesian information criteria (`measure='bic'`) or, for `SARIMAX` and `UnobservedComponents`, the Hannan–Quinn information criterion (`measure='hqic'`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "res_sarimax_bic = sm.tsa.automatic.auto_sarimax(\n",
    "    dta.cpi, measure='bic', d=2, stepwise=True, fit_kwargs={'maxiter': 100})\n",
    "res_sarimax_hqic = sm.tsa.automatic.auto_sarimax(\n",
    "    dta.cpi, measure='hqic', d=2, stepwise=True, fit_kwargs={'maxiter': 100})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we will see below when we use `specification_search` to examine the results from all considered models, the BIC and HQIC select a different model than the AIC does, although the model selected by the BIC and HQIC is the same.\n",
    "\n",
    "In this case, the AIC selects a model with a slightly higher log-likelihood but includes three more parameters than the model selected by the BIC or HQIC."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AIC-selected model:  llf=-230.69, k_params=7\n",
      "BIC-selected model:  llf=-234.80, k_params=4\n",
      "HQIC-selected model: llf=-234.80, k_params=4\n"
     ]
    }
   ],
   "source": [
    "print('AIC-selected model:  llf=%.2f, k_params=%d' % (res_sarimax.llf, len(res_sarimax.params)))\n",
    "print('BIC-selected model:  llf=%.2f, k_params=%d' % (res_sarimax_bic.llf, len(res_sarimax_bic.params)))\n",
    "print('HQIC-selected model: llf=%.2f, k_params=%d' % (res_sarimax_hqic.llf, len(res_sarimax_hqic.params)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Specification search\n",
    "\n",
    "Alternatively, we may want to automatically perform the selection search, but then manually inspect the results before selecting a final model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                          Specification Search Results                          \n",
      "================================================================================\n",
      "Dep. Variable:                      cpi   No. Observations:                  203\n",
      "Model:             ExponentialSmoothing   No. Specifications:                  3\n",
      "Date:                  Sat, 08 Sep 2018   Measure                            AIC\n",
      "Time:                          23:37:55   Minimum AIC                    -80.653\n",
      "Sample:                      03-31-1959                                         \n",
      "                           - 09-30-2009                                         \n",
      "                     Evaluated models                     \n",
      "==========================================================\n",
      "   trend damped   aic    bic   k_params warnings/exception\n",
      "----------------------------------------------------------\n",
      "0   add  False  -80.653 -67.40    4             No        \n",
      "1   add   True  -78.970 -62.40    5             No        \n",
      "2   None False  105.258 111.88    2             No        \n",
      "----------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# Exponential smoothing\n",
    "search_res_es = sm.tsa.automatic.exponentialsmoothing.specification_search(dta.cpi)\n",
    "print(search_res_es.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                         Specification Search Results                         \n",
      "==============================================================================\n",
      "Dep. Variable:                    cpi   No. Observations:                  203\n",
      "Model:                        SARIMAX   No. Specifications:                  9\n",
      "Date:                Sat, 08 Sep 2018   Measure                            AIC\n",
      "Time:                        23:37:57   Minimum AIC                    475.379\n",
      "Sample:                    03-31-1959                                         \n",
      "                         - 09-30-2009                                         \n",
      "                          Evaluated models                          \n",
      "====================================================================\n",
      "     order   trend   aic    bic    hqic  k_params warnings/exception\n",
      "--------------------------------------------------------------------\n",
      "0  (3, 2, 3)   n   475.379 498.50 484.74    7             No        \n",
      "1  (1, 2, 2)   n   477.600 490.81 482.95    4             No        \n",
      "2  (2, 2, 3)   n   477.805 497.62 485.82    6             No        \n",
      "3  (2, 2, 2)   n   479.129 495.65 485.81    5             No        \n",
      "4  (3, 2, 2)   n   480.530 500.35 488.55    6            Yes        \n",
      "5  (1, 2, 1)   n   486.399 496.31 490.41    3             No        \n",
      "6  (2, 2, 1)   n   487.878 501.09 493.22    4             No        \n",
      "7  (1, 2, 0)   n   539.482 546.09 542.16    2             No        \n",
      "8  (0, 2, 0)   n   567.417 570.72 568.75    1             No        \n",
      "--------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# SARIMAX\n",
    "search_res_sarimax = sm.tsa.automatic.sarimax.specification_search(\n",
    "    dta.cpi, d=2, stepwise=True, fit_kwargs={'maxiter': 100})\n",
    "print(search_res_sarimax.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                          Specification Search Results                          \n",
      "================================================================================\n",
      "Dep. Variable:                      cpi   No. Observations:                  203\n",
      "Model:             UnobservedComponents   No. Specifications:                  3\n",
      "Date:                  Sat, 08 Sep 2018   Measure                            AIC\n",
      "Time:                          23:37:58   Minimum AIC                    491.928\n",
      "Sample:                      03-31-1959                                         \n",
      "                           - 09-30-2009                                         \n",
      "                            Evaluated models                           \n",
      "=======================================================================\n",
      "         level          aic    bic    hqic  k_params warnings/exception\n",
      "-----------------------------------------------------------------------\n",
      "0  local linear trend 491.928 501.84 495.94    3             No        \n",
      "1     smooth trend    526.256 532.86 528.93    2             No        \n",
      "2     random trend    567.417 570.72 568.75    1             No        \n",
      "-----------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# Unobserved components\n",
    "search_res_uc = sm.tsa.automatic.structural.specification_search(\n",
    "    dta.cpi, level='I(2)')\n",
    "print(search_res_uc.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Saving results\n",
    "\n",
    "By default, the `Results` objects constructed during the `selection_search` call are all stored (and accessible via e.g. `search_res_uc[1].results`). To save memory, however, the `remove_data` method has been called; this means that much of the usual output - including out-of-sample forecasts - is unavailable.\n",
    "\n",
    "If it is known that multiple results objects will be used, one can pass the argument `store_results_data=True`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "stored_search_res_sarimax = sm.tsa.automatic.sarimax.specification_search(\n",
    "    dta.cpi, d=2, stepwise=True, fit_kwargs={'maxiter': 100}, store_results_data=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can then use the stored results to compare the MSE of the various considered models:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean squared error (in-sample):\n",
      " - SARIMAX(endog, order=(3, 2, 3), trend=\"n\"):   0.571\n",
      " - SARIMAX(endog, order=(1, 2, 2), trend=\"n\"):   0.600\n",
      " - SARIMAX(endog, order=(2, 2, 3), trend=\"n\"):   0.585\n",
      " - SARIMAX(endog, order=(2, 2, 2), trend=\"n\"):   0.599\n",
      " - SARIMAX(endog, order=(3, 2, 2), trend=\"n\"):   0.597\n",
      " - SARIMAX(endog, order=(1, 2, 1), trend=\"n\"):   0.634\n",
      " - SARIMAX(endog, order=(2, 2, 1), trend=\"n\"):   0.633\n",
      " - SARIMAX(endog, order=(1, 2, 0), trend=\"n\"):   0.840\n",
      " - SARIMAX(endog, order=(0, 2, 0), trend=\"n\"):   0.975\n"
     ]
    }
   ],
   "source": [
    "print('Mean squared error (in-sample):')\n",
    "for key, spec in stored_search_res_sarimax.items():\n",
    "    print(' - %s:   %.3f' % (str(spec.specification), np.mean(spec.results.resid[2:]**2)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Or use the stored results to produce a variety of out-of-sample forecasts:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(13, 5))\n",
    "\n",
    "dta.cpi.loc['2006':].plot(ax=ax, label='Data')\n",
    "for key, spec in stored_search_res_sarimax.items():\n",
    "    spec.results.forecast(10).plot(ax=ax, label=str(spec.specification))\n",
    "\n",
    "ax.set(title='Forecasts from SARIMAX models considered in specification search')\n",
    "ax.legend();"
   ]
  }
 ],
 "metadata": {
  "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}