{
  "cells": [
    {
      "cell_type": "code",
      "metadata": {},
      "source": [
        "#%%\n",
        "\"\"\"File 04validation.py\n",
        "\n",
        ":author: Michel Bierlaire, EPFL\n",
        ":date: Thu Jun  4 17:55:27 2020\n",
        "\n",
        " Example of the out-of-sample validation of a logit model.\n",
        " Three alternatives: Train, Car and Swissmetro\n",
        " SP data\n",
        "\"\"\"\n",
        "\n",
        "import pandas as pd\n",
        "import biogeme.database as db\n",
        "import biogeme.biogeme as bio\n",
        "import biogeme.models as models\n",
        "from biogeme.expressions import Beta\n",
        "\n",
        "# Read the data\n",
        "df = pd.read_csv('swissmetro.dat', '\\t')\n",
        "database = db.Database('swissmetro', df)\n",
        "\n",
        "# The following statement allows you to use the names of the variable\n",
        "# as Python variable.\n",
        "globals().update(database.variables)\n",
        "\n",
        "# Removing some observations\n",
        "exclude = ((PURPOSE != 1) * (PURPOSE != 3) + (CHOICE == 0)) > 0\n",
        "database.remove(exclude)\n",
        "\n",
        "# Parameters to be estimated\n",
        "ASC_CAR = Beta('ASC_CAR', 0, None, None, 0)\n",
        "ASC_TRAIN = Beta('ASC_TRAIN', 0, None, None, 0)\n",
        "ASC_SM = Beta('ASC_SM', 0, None, None, 1)\n",
        "B_TIME = Beta('B_TIME', 0, None, None, 0)\n",
        "B_COST = Beta('B_COST', 0, None, None, 0)\n",
        "\n",
        "\n",
        "# Definition of new variables\n",
        "SM_COST = SM_CO * (GA == 0)\n",
        "TRAIN_COST = TRAIN_CO * (GA == 0)\n",
        "CAR_AV_SP = CAR_AV * (SP != 0)\n",
        "TRAIN_AV_SP = TRAIN_AV * (SP != 0)\n",
        "TRAIN_TT_SCALED = TRAIN_TT / 100.0\n",
        "TRAIN_COST_SCALED = TRAIN_COST / 100\n",
        "SM_TT_SCALED = SM_TT / 100.0\n",
        "SM_COST_SCALED = SM_COST / 100\n",
        "CAR_TT_SCALED = CAR_TT / 100\n",
        "CAR_CO_SCALED = CAR_CO / 100\n",
        "\n",
        "# Definition of the utility functions\n",
        "V1 = ASC_TRAIN + \\\n",
        "     B_TIME * TRAIN_TT_SCALED + \\\n",
        "     B_COST * TRAIN_COST_SCALED\n",
        "V2 = ASC_SM + \\\n",
        "     B_TIME * SM_TT_SCALED + \\\n",
        "     B_COST * SM_COST_SCALED\n",
        "V3 = ASC_CAR + \\\n",
        "     B_TIME * CAR_TT_SCALED + \\\n",
        "     B_COST * CAR_CO_SCALED\n",
        "\n",
        "# Associate utility functions with the numbering of alternatives\n",
        "V = {1: V1,\n",
        "     2: V2,\n",
        "     3: V3}\n",
        "\n",
        "# Associate the availability conditions with the alternatives\n",
        "av = {1: TRAIN_AV_SP,\n",
        "      2: SM_AV,\n",
        "      3: CAR_AV_SP}\n",
        "\n",
        "# Definition of the model. This is the contribution of each\n",
        "# observation to the log likelihood function.\n",
        "logprob = models.loglogit(V, av, CHOICE)\n",
        "\n",
        "# Create the Biogeme object\n",
        "biogeme = bio.BIOGEME(database, logprob)\n",
        "biogeme.modelName = '04validation'\n",
        "\n",
        "# Estimate the parameters\n",
        "results = biogeme.estimate()\n",
        "\n",
        "# The validation consists in organizing the data into several slices\n",
        "# of about the same size, randomly defined. Each slice is considered\n",
        "# as a validation dataset. The model is then re-estimated using all\n",
        "# the data except the slice, and the estimated model is applied on the\n",
        "# validation set (i.e. the slice). The value of the log likelihood for\n",
        "# each observation in the validation set is reported in a\n",
        "# dataframe. As this is done for each slice, the output is a list of\n",
        "# dataframes, each corresponding to one of these exercises.\n",
        "\n",
        "validation_results = biogeme.validate(results)\n",
        "\n",
        "for slide in validation_results:\n",
        "    print(f'Log likelihood for {slide.shape[0]} validation data: {slide[\"Loglikelihood\"].sum()}')\n"
      ],
      "outputs": [],
      "execution_count": null
    }
  ],
  "metadata": {
    "anaconda-cloud": {},
    "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": "3.6.1"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 1
}