docs/tutorials/regression_prior.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regression Priors in DLT"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook demonstrates usage of priors in the regression analysis. The *iclaims* data will be used in demo purpose. Examples include\n",
    "\n",
    "1. regression with default setting\n",
    "2. regression with bounded priors for regression coefficients \n",
    "\n",
    "Generally speaking, regression coefficients are more robust under full Bayesian sampling and estimation.  The default setting `estimator='stan-mcmc'` will be used in this tutorial."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-28T21:49:38.157658Z",
     "start_time": "2022-04-28T21:49:35.620843Z"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "import orbit\n",
    "from orbit.utils.dataset import load_iclaims\n",
    "from orbit.models import DLT\n",
    "from orbit.diagnostics.plot import plot_predicted_data\n",
    "from orbit.constants.palette import OrbitPalette"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-28T21:49:38.163035Z",
     "start_time": "2022-04-28T21:49:38.160107Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.1.3\n"
     ]
    }
   ],
   "source": [
    "print(orbit.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "toc-hr-collapsed": false
   },
   "source": [
    "## US Weekly Initial Claims"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Recall the *iclaims* dataset by previous section. In order to use this data to nowcast the US unemployment claims during COVID-19 period, the dataset is extended to Jan 2021 and the [S&P 500 (^GSPC)](https://finance.yahoo.com/quote/%5EGSPC/history?period1=1264032000&period2=1611187200&interval=1wk&filter=history&frequency=1wk&includeAdjustedClose=true) and [VIX](https://finance.yahoo.com/quote/%5EVIX/history?p=%5EVIX) Index historical data are attached for the same period.\n",
    "\n",
    "The data is standardized and log-transformed for the model fitting purpose."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:48:33.076267Z",
     "start_time": "2022-04-07T18:48:31.033862Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "week              datetime64[ns]\n",
       "claims                   float64\n",
       "trend.unemploy           float64\n",
       "trend.job                float64\n",
       "sp500                    float64\n",
       "vix                      float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# load data\n",
    "df = load_iclaims(end_date='2021-01-03')\n",
    "df = df[['week', 'claims', 'trend.unemploy', 'trend.job', 'sp500', 'vix']]\n",
    "df = df[1:].reset_index(drop=True)\n",
    "\n",
    "date_col = 'week'\n",
    "response_col = 'claims'\n",
    "df.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:48:33.091950Z",
     "start_time": "2022-04-07T18:48:33.079051Z"
    }
   },
   "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>week</th>\n",
       "      <th>claims</th>\n",
       "      <th>trend.unemploy</th>\n",
       "      <th>trend.job</th>\n",
       "      <th>sp500</th>\n",
       "      <th>vix</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2010-01-10</td>\n",
       "      <td>13.624218</td>\n",
       "      <td>0.016351</td>\n",
       "      <td>0.181862</td>\n",
       "      <td>-0.550891</td>\n",
       "      <td>0.069878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>13.398741</td>\n",
       "      <td>0.032611</td>\n",
       "      <td>0.130569</td>\n",
       "      <td>-0.590640</td>\n",
       "      <td>0.491772</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2010-01-24</td>\n",
       "      <td>13.137549</td>\n",
       "      <td>-0.000179</td>\n",
       "      <td>0.119987</td>\n",
       "      <td>-0.607162</td>\n",
       "      <td>0.388078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2010-01-31</td>\n",
       "      <td>13.196760</td>\n",
       "      <td>-0.069172</td>\n",
       "      <td>0.087552</td>\n",
       "      <td>-0.614339</td>\n",
       "      <td>0.446838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2010-02-07</td>\n",
       "      <td>13.146984</td>\n",
       "      <td>-0.182500</td>\n",
       "      <td>0.019344</td>\n",
       "      <td>-0.605636</td>\n",
       "      <td>0.308205</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        week     claims  trend.unemploy  trend.job     sp500       vix\n",
       "0 2010-01-10  13.624218        0.016351   0.181862 -0.550891  0.069878\n",
       "1 2010-01-17  13.398741        0.032611   0.130569 -0.590640  0.491772\n",
       "2 2010-01-24  13.137549       -0.000179   0.119987 -0.607162  0.388078\n",
       "3 2010-01-31  13.196760       -0.069172   0.087552 -0.614339  0.446838\n",
       "4 2010-02-07  13.146984       -0.182500   0.019344 -0.605636  0.308205"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see form the plot below, there are seasonality, trend, and as well as a huge change point due the impact of COVID-19."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:48:33.656211Z",
     "start_time": "2022-04-07T18:48:33.093772Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'S&P500')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(2, 2,figsize=(20,8))\n",
    "axs[0, 0].plot(df['week'], df['claims'])\n",
    "axs[0, 0].set_title('Unemployment Claims')\n",
    "axs[0, 1].plot(df['week'], df['trend.unemploy'], 'tab:orange')\n",
    "axs[0, 1].set_title('Google trend - unemploy')\n",
    "axs[1, 0].plot(df['week'], df['vix'], 'tab:green')\n",
    "axs[1, 0].set_title('VIX')\n",
    "axs[1, 1].plot(df['week'], df['sp500'], 'tab:red')\n",
    "axs[1, 1].set_title('S&P500')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:48:33.662455Z",
     "start_time": "2022-04-07T18:48:33.657947Z"
    }
   },
   "outputs": [],
   "source": [
    "# using relatively updated data\n",
    "df[['sp500']] = df[['sp500']].diff()\n",
    "df = df[1:].reset_index(drop=True)\n",
    "\n",
    "test_size = 12\n",
    "train_df = df[:-test_size]\n",
    "test_df = df[-test_size:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Naive Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we will use DLT models to compare the model performance with vs. without regression."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:48:46.681927Z",
     "start_time": "2022-04-07T18:48:33.665046Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:orbit:Sampling (PyStan) with chains: 4, cores: 8, temperature: 1.000, warmups (per chain): 1000 and samples(per chain): 25.\n",
      "WARNING:pystan:Maximum (flat) parameter count (1000) exceeded: skipping diagnostic tests for n_eff and Rhat.\n",
      "To run all diagnostics call pystan.check_hmc_diagnostics(fit)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 152 ms, sys: 108 ms, total: 260 ms\n",
      "Wall time: 7.01 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "dlt = DLT(\n",
    "    response_col=response_col,\n",
    "    date_col=date_col,\n",
    "    seasonality=52,\n",
    "    seed=8888,\n",
    "    num_warmup=4000,\n",
    ")\n",
    "dlt.fit(df=train_df)\n",
    "predicted_df = dlt.predict(df=test_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "toc-hr-collapsed": false
   },
   "source": [
    "### DLT With Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The regressor columns can be supplied via argument `regressor_col`.  Recall the regression formula in **DLT**:\n",
    "\n",
    "$$\n",
    "\\hat{y}_t =\\mu_t + s_t + r_t \\\\\n",
    "r_t = \\sum_{j}\\beta_j x_{jt} \\\\\n",
    "\\beta_j ~\\sim \\mathcal{N}(\\mu_j, \\sigma_j^2)\n",
    "$$\n",
    "\n",
    "By default, $\\mu_j = 0$ and $\\sigma_j = 1$.  In addition, we can set a *sign* constraint for each coefficient $\\beta_j$.  This is can be done by supplying the `regressor_sign` as a list where elements are in one of followings:\n",
    "\n",
    "* '=': $\\beta_j ~\\sim \\mathcal{N}(0, \\sigma_j^2)$  i.e. $\\beta_j \\in (-\\inf, \\inf)$\n",
    "* '+': $\\beta_j ~\\sim \\mathcal{N}^+(0, \\sigma_j^2)$  i.e. $\\beta_j \\in [0, \\inf)$\n",
    "* '-': $\\beta_j ~\\sim \\mathcal{N}^-(0, \\sigma_j^2)$  i.e. $\\beta_j \\in (-\\inf, 0]$\n",
    "\n",
    "Based on some intuition, it's reasonable to assume search terms such as \"unemployment\", \"filling\" and **VIX** index to be positively correlated and stock index such as **SP500** to be negatively correlated to the outcome.  Then we will leave whatever unspecified as a regular regressor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:49:00.202316Z",
     "start_time": "2022-04-07T18:48:46.685743Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:orbit:Sampling (PyStan) with chains: 4, cores: 8, temperature: 1.000, warmups (per chain): 1000 and samples(per chain): 25.\n",
      "WARNING:pystan:Maximum (flat) parameter count (1000) exceeded: skipping diagnostic tests for n_eff and Rhat.\n",
      "To run all diagnostics call pystan.check_hmc_diagnostics(fit)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 81.1 ms, sys: 85.6 ms, total: 167 ms\n",
      "Wall time: 6.96 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "dlt_reg = DLT(\n",
    "    response_col=response_col, \n",
    "    date_col=date_col,\n",
    "    regressor_col=['trend.unemploy', 'trend.job', 'sp500', 'vix'],\n",
    "    seasonality=52,\n",
    "    seed=8888,\n",
    "    num_warmup=4000,\n",
    ")\n",
    "dlt_reg.fit(df=train_df)\n",
    "predicted_df_reg = dlt_reg.predict(test_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The estimated regressor coefficients can be retrieved via `.get_regression_coefs()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:49:00.217532Z",
     "start_time": "2022-04-07T18:49:00.204563Z"
    }
   },
   "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>regressor</th>\n",
       "      <th>regressor_sign</th>\n",
       "      <th>coefficient</th>\n",
       "      <th>coefficient_lower</th>\n",
       "      <th>coefficient_upper</th>\n",
       "      <th>Pr(coef &gt;= 0)</th>\n",
       "      <th>Pr(coef &lt; 0)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>trend.unemploy</td>\n",
       "      <td>Regular</td>\n",
       "      <td>0.080230</td>\n",
       "      <td>0.048558</td>\n",
       "      <td>0.106878</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>trend.job</td>\n",
       "      <td>Regular</td>\n",
       "      <td>-0.039983</td>\n",
       "      <td>-0.083835</td>\n",
       "      <td>0.009523</td>\n",
       "      <td>0.13</td>\n",
       "      <td>0.87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>sp500</td>\n",
       "      <td>Regular</td>\n",
       "      <td>-0.010453</td>\n",
       "      <td>-0.202477</td>\n",
       "      <td>0.215410</td>\n",
       "      <td>0.42</td>\n",
       "      <td>0.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>vix</td>\n",
       "      <td>Regular</td>\n",
       "      <td>0.010462</td>\n",
       "      <td>-0.024078</td>\n",
       "      <td>0.036222</td>\n",
       "      <td>0.71</td>\n",
       "      <td>0.29</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        regressor regressor_sign  coefficient  coefficient_lower  \\\n",
       "0  trend.unemploy        Regular     0.080230           0.048558   \n",
       "1       trend.job        Regular    -0.039983          -0.083835   \n",
       "2           sp500        Regular    -0.010453          -0.202477   \n",
       "3             vix        Regular     0.010462          -0.024078   \n",
       "\n",
       "   coefficient_upper  Pr(coef >= 0)  Pr(coef < 0)  \n",
       "0           0.106878           1.00          0.00  \n",
       "1           0.009523           0.13          0.87  \n",
       "2           0.215410           0.42          0.58  \n",
       "3           0.036222           0.71          0.29  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dlt_reg.get_regression_coefs()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Regression with Informative Priors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Due to various reasons, users may obtain further knowledge on some of the regressors or they want to propose different regularization on different regressors. These informative priors basically means to replace the defaults ($\\mu$, $\\sigma$) mentioned previously. In orbit, this process is done via the arguments `regressor_beta_prior` and `regressor_sigma_prior`. These two lists should be of the same length as `regressor_col`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In addition, we can set a *sign* constraint for each coefficient $\\beta_j$.  This is can be done by supplying the `regressor_sign` as a list where elements are in one of followings:\n",
    "\n",
    "* '=': $\\beta_j ~\\sim \\mathcal{N}(0, \\sigma_j^2)$  i.e. $\\beta_j \\in (-\\inf, \\inf)$\n",
    "* '+': $\\beta_j ~\\sim \\mathcal{N}^+(0, \\sigma_j^2)$  i.e. $\\beta_j \\in [0, \\inf)$\n",
    "* '-': $\\beta_j ~\\sim \\mathcal{N}^-(0, \\sigma_j^2)$  i.e. $\\beta_j \\in (-\\inf, 0]$\n",
    "\n",
    "Based on intuition, it's reasonable to assume search terms such as \"unemployment\", \"filling\" and **VIX** index to be positively correlated (`+` sign is used in this case) and upward shock of **SP500** (`-` sign) to be negatively correlated to the outcome. Otherwise, an unbounded coefficient can be used (`=` sign).\n",
    "\n",
    "Furthermore, regressors such as search queries may have more direct impact than stock marker indices.  Hence, a smaller $\\sigma$ is considered."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:49:21.970562Z",
     "start_time": "2022-04-07T18:49:00.219185Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:orbit:Sampling (PyStan) with chains: 4, cores: 8, temperature: 1.000, warmups (per chain): 1000 and samples(per chain): 250.\n",
      "WARNING:pystan:Maximum (flat) parameter count (1000) exceeded: skipping diagnostic tests for n_eff and Rhat.\n",
      "To run all diagnostics call pystan.check_hmc_diagnostics(fit)\n"
     ]
    }
   ],
   "source": [
    "dlt_reg_adjust = DLT(\n",
    "    response_col=response_col,\n",
    "    date_col=date_col,\n",
    "    regressor_col=['trend.unemploy', 'trend.job', 'sp500','vix'],\n",
    "    regressor_sign=['+','=','-','+'],\n",
    "    regressor_sigma_prior=[0.3, 0.1, 0.05, 0.1],\n",
    "    num_warmup=4000,\n",
    "    num_sample=1000,\n",
    "    estimator='stan-mcmc',\n",
    "    seed=2022,\n",
    ")\n",
    "dlt_reg_adjust.fit(df=train_df)\n",
    "predicted_df_reg_adjust = dlt_reg_adjust.predict(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:49:21.986924Z",
     "start_time": "2022-04-07T18:49:21.972519Z"
    }
   },
   "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>regressor</th>\n",
       "      <th>regressor_sign</th>\n",
       "      <th>coefficient</th>\n",
       "      <th>coefficient_lower</th>\n",
       "      <th>coefficient_upper</th>\n",
       "      <th>Pr(coef &gt;= 0)</th>\n",
       "      <th>Pr(coef &lt; 0)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>trend.unemploy</td>\n",
       "      <td>Positive</td>\n",
       "      <td>0.132037</td>\n",
       "      <td>0.078566</td>\n",
       "      <td>0.210839</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>vix</td>\n",
       "      <td>Positive</td>\n",
       "      <td>0.019541</td>\n",
       "      <td>0.001662</td>\n",
       "      <td>0.054155</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>sp500</td>\n",
       "      <td>Negative</td>\n",
       "      <td>-0.030996</td>\n",
       "      <td>-0.095246</td>\n",
       "      <td>-0.003587</td>\n",
       "      <td>0.000</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>trend.job</td>\n",
       "      <td>Regular</td>\n",
       "      <td>-0.011236</td>\n",
       "      <td>-0.080350</td>\n",
       "      <td>0.060355</td>\n",
       "      <td>0.386</td>\n",
       "      <td>0.614</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        regressor regressor_sign  coefficient  coefficient_lower  \\\n",
       "0  trend.unemploy       Positive     0.132037           0.078566   \n",
       "1             vix       Positive     0.019541           0.001662   \n",
       "2           sp500       Negative    -0.030996          -0.095246   \n",
       "3       trend.job        Regular    -0.011236          -0.080350   \n",
       "\n",
       "   coefficient_upper  Pr(coef >= 0)  Pr(coef < 0)  \n",
       "0           0.210839          1.000         0.000  \n",
       "1           0.054155          1.000         0.000  \n",
       "2          -0.003587          0.000         1.000  \n",
       "3           0.060355          0.386         0.614  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dlt_reg_adjust.get_regression_coefs()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's compare the holdout performance by using the built-in function `smape()` ."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-04-07T18:49:21.994703Z",
     "start_time": "2022-04-07T18:49:21.988876Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------Mean Absolute Error Summary----------------\n",
      "Naive Model: 0.249\n",
      "Regression Model: 0.249\n",
      "Refined Regression Model: 0.097\n"
     ]
    }
   ],
   "source": [
    "def mae(x, y):\n",
    "    return np.mean(np.abs(x - y))\n",
    "\n",
    "naive_mae = mae(predicted_df['prediction'].values, test_df['claims'].values)\n",
    "reg_mae = mae(predicted_df_reg['prediction'].values, test_df['claims'].values)\n",
    "reg_adjust_mae = mae(predicted_df_reg_adjust['prediction'].values, test_df['claims'].values)\n",
    "\n",
    "print(\"----------------Mean Absolute Error Summary----------------\")\n",
    "print(\"Naive Model: {:.3f}\\nRegression Model: {:.3f}\\nRefined Regression Model: {:.3f}\".format(\n",
    "    naive_mae, reg_mae, reg_adjust_mae\n",
    "))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary\n",
    "\n",
    "This demo showcases a use case in nowcasting. Although this may not be applicable in real-time forecasting, it mainly introduces the regression analysis with time-series modeling in `Orbit`. For people who have concerns on the forecastability, one can consider introducing lag on regressors.\n",
    "\n",
    "Also, `Orbit` allows informative priors where sometime can be useful in combining multiple source of insights together."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "orbit38",
   "language": "python",
   "name": "orbit38"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.12"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}