{ "cells": [ { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "

Example-Dependent Cost-Sensitive Credit Scoring using CostCla

" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "-" } }, "source": [ "
\n", "\n", "

Alejandro Correa Bahnsen

\n", "\n", "

\n", "\n", "

\n", "\n", "
\n", "\n", "
\n", "\n", "
\n", "\n", "

PyData Berlin, May 2015

\n", "
" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "

About Me

" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "skip" } }, "source": [ "%%html\n", "" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "###A brief bio:\n", "\n", "* Last year (month) PhD Student at Luxembourg University\n", "* Work part time a fraud data scientist at CETREL a SIX Company\n", "* Worked for +5 years as a data scientist at GE Money and Scotiabank\n", "* Previously, six sigma intern at Dow Chemical\n", "* Bachelor in Industrial Engineering and Master in Financial Engineering\n", "* Organizer of Data Science Luxembourg and recently of Big Data Science Bog \n", "* Sport addict, love to swim, play tennis, squash, and volleyball, among others.\n", "\n", "

\n", "\n", "\n", " \n", " \n", "\n", "
\n", "\n", " \n", " \n", " \n", " \n", "\n", "al.bahnsen@gmail.com \n", "
\n", " \n", "\n", " \n", " \n", " \n", "\n", "\n", "\n", "http://github.com/albahnsen \n", "\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "http://linkedin.com/in/albahnsen \n", "\n", "
\n", "\n", "\n", " \n", " \n", " \n", "\n", " \n", "@albahnsen \n", "\n", "
" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "# Agenda\n", "\n", "* Quick Intro to Credit Scoring\n", "* Example of Credit Scoring\n", "* Financial Evaluation of a Credit Scorecard\n", "* Example-Dependent Classification\n", "* CostCla Library\n", "* Conclusion and Future Work" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "

Credit Scoring

" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "#To whom would you grant a loan?" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "| | |\n", "|:-:|:-:|\n", "| Just fund a bank | Just quit college |\n", "\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Nice guess!\n", "\n", "| | |\n", "|:-:|:-:|\n", "| Biggest Ponzi scheme | Now a Billionaire |\n", "\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Credit Scoring" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Mitigate the impact of **credit risk** and make more objective\n", "and accurate decisions" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Estimate the **risk of a customer defaulting** his contracted\n", "financial obligation if a loan is granted, based on past\n", "experiences" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Different machine learning methods are used in practice, and in the\n", "literature: logistic regression, neural networks, discriminant\n", "analysis, genetic programing, decision trees, random forests among others" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Credit Scoring" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "-" } }, "source": [ "Formally, a credit score is a statistical model that allows the estimation of the probability of a customer $i$ defaulting a contracted debt ($y_i=1$)\n", "\n", " $$\\hat p_i=P(y_i=1|\\mathbf{x}_i)$$ " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "

Example: Kaggle Credit Competition

" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "\n", "
\n", "\n", "
\n", "\n", "Improve on the state of the art in credit scoring by predicting the probability that somebody will experience financial distress in the next two years.\n", "\n", "https://www.kaggle.com/c/GiveMeSomeCredit" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Load dataset from CostCla package" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "collapsed": false, "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "import pandas as pd\n", "import numpy as np" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "collapsed": false, "slideshow": { "slide_type": "fragment" } }, "outputs": [], "source": [ "from costcla.datasets import load_creditscoring1\n", "data = load_creditscoring1()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "### Data file" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "collapsed": false, "slideshow": { "slide_type": "-" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "['target_names', 'cost_mat', 'name', 'DESCR', 'feature_names', 'data', 'target']\n", "Number of examples 112915\n" ] } ], "source": [ "print data.keys()\n", "print 'Number of examples ', data.target.shape[0]" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Class Label" ] }, { "cell_type": "code", "execution_count": 14, "metadata": { "collapsed": false, "scrolled": true }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Frequency Percentage\n", "Negative (Good Customers) 105299 0.932551\n", "Positive (Bad Customers) 7616 0.067449\n" ] } ], "source": [ "target = pd.DataFrame(pd.Series(data.target).value_counts(), columns=('Frequency',))\n", "target['Percentage'] = target['Frequency'] / target['Frequency'].sum()\n", "target.index = ['Negative (Good Customers)', 'Positive (Bad Customers)']\n", "print target" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Features" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Features
0RevolvingUtilizationOfUnsecuredLines
1age
2NumberOfTime30-59DaysPastDueNotWorse
3DebtRatio
4MonthlyIncome
5NumberOfOpenCreditLinesAndLoans
6NumberOfTimes90DaysLate
7NumberRealEstateLoansOrLines
8NumberOfTime60-89DaysPastDueNotWorse
9NumberOfDependents
\n", "
" ], "text/plain": [ " Features\n", "0 RevolvingUtilizationOfUnsecuredLines\n", "1 age\n", "2 NumberOfTime30-59DaysPastDueNotWorse\n", "3 DebtRatio\n", "4 MonthlyIncome\n", "5 NumberOfOpenCreditLinesAndLoans\n", "6 NumberOfTimes90DaysLate\n", "7 NumberRealEstateLoansOrLines\n", "8 NumberOfTime60-89DaysPastDueNotWorse\n", "9 NumberOfDependents" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pd.DataFrame(data.feature_names, columns=('Features',))" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Credit scoring as a classification problem" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "-" } }, "source": [ "### Split in training and testing" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "collapsed": false, "slideshow": { "slide_type": "-" } }, "outputs": [], "source": [ "from sklearn.cross_validation import train_test_split\n", "X_train, X_test, y_train, y_test, cost_mat_train, cost_mat_test = \\\n", "train_test_split(data.data, data.target, data.cost_mat)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Credit scoring as a classification problem" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "-" } }, "source": [ "### Fit models" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "collapsed": false, "slideshow": { "slide_type": "-" } }, "outputs": [], "source": [ "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.tree import DecisionTreeClassifier\n", "\n", "classifiers = {\"RF\": {\"f\": RandomForestClassifier()},\n", " \"DT\": {\"f\": DecisionTreeClassifier()},\n", " \"LR\": {\"f\": LogisticRegression()}}\n", "\n", "# Fit the classifiers using the training dataset\n", "for model in classifiers.keys():\n", " classifiers[model][\"f\"].fit(X_train, y_train)\n", " classifiers[model][\"c\"] = classifiers[model][\"f\"].predict(X_test)\n", " classifiers[model][\"p\"] = classifiers[model][\"f\"].predict_proba(X_test)\n", " classifiers[model][\"p_train\"] = classifiers[model][\"f\"].predict_proba(X_train)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Models performance" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "-" } }, "source": [ "### Evaluate metrics and plot results" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "collapsed": false, "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "%matplotlib inline\n", "import matplotlib.pyplot as plt\n", "from IPython.core.pylabtools import figsize\n", "import seaborn as sns\n", "figsize(12, 8)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "collapsed": false, "slideshow": { "slide_type": "-" } }, "outputs": [], "source": [ "from sklearn.metrics import f1_score, precision_score, recall_score, accuracy_score\n", "\n", "measures = {\"F1Score\": f1_score, \"Precision\": precision_score, \n", " \"Recall\": recall_score, \"Accuracy\": accuracy_score}\n", "\n", "results = pd.DataFrame(columns=measures.keys())\n", "\n", "for model in classifiers.keys():\n", " results.loc[model] = [measures[measure](y_test, classifiers[model][\"c\"]) for measure in measures.keys()]" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "collapsed": false, "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "def fig1():\n", " plt.figure()\n", " l = plt.plot(range(results.shape[0]), results, \"-o\", linewidth=7, markersize=15)\n", " plt.legend(iter(l), results.columns.tolist(), loc='center left', bbox_to_anchor=(1, 0.5),fontsize=22)\n", " plt.xlim([-0.25, results.shape[0]-1+.25])\n", " plt.xticks(range(results.shape[0]), results.index) \n", " plt.tick_params(labelsize=22)\n", " plt.show()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Models performance" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "collapsed": false, "slideshow": { "slide_type": "-" } }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4QAAAHzCAYAAAB49tNrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig1()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Models performance" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- None of these measures takes into account the **business and economical realities** that take place in credit scoring. " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Costs that the financial institution had incurred to acquire customers, or the **expected profit** due to a particular client, are not considered in the evaluation of the different models. " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "

Financial Evaluation of a Credit Scorecard

" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Motivation\n", "\n", "- Typically, a credit risk model is evaluated using standard **cost-insensitive measures**.\n", "- However, in practice, the cost associated with **approving a bad customer** (False Negative) is quite different from the cost associated with **declining a good customer** (False Positive).\n", "- Furthermore, the costs are **not constant** among customers. " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Cost Matrix\n", "\n", " \n", "| \t| Actual Positive ($y_i=1$) \t| Actual Negative \t($y_i=0$)|\n", "|---\t|:-:\t|:-:\t|\n", "| Pred. Positive ($c_i=1$)\t| $C_{TP_i}=0$\t| $C_{FP_i}=r_i+C^a_{FP}$ \t|\n", "| Pred. Negative ($c_i=0$) \t| $C_{FN_i}=Cl_i \\cdot L_{gd}$\t| $C_{TN_i}=0$\t|\n", "\n", "Where:\n", "\n", "- $C_{FN_i}$ = losses if the customer $i$ defaults\n", "- $Cl_i$ is the credi line of customer $i$\n", "- $L_{gd}$ is the loss given default. Percentage of loss over the total credit line when the customer defaulted" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Cost Matrix\n", "\n", "- $C_{FP_i}=r_i+C^a_{FP}$ \n", "- $r_i$ is the loss in profit by rejecting what would have been a good customer.\n", "- $C^a_{FP}$ is related to the assumption that the financial institution will not keep the money of the declined customer idle, but instead it will give\n", "a loan to an alternative customer.\n", "\n", "For more info see [Correa Bahnsen et al., 2014] " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Parameters for the Kaggle Credit Database\n", "\n", "Assuming the database belong to an average European financial institution, we find the different parameters needed to calculate the cost measure\n", "\n", "| Parameter \t| Value |\n", "|---\t|:-:\t|\n", "|Interest rate ($int_r$) | 4.79% |\n", "| Cost of funds ($int_{cf}$) | 2.94% |\n", "| Term ($l$) in months | 24 |\n", "| Loss given default ($L_{gd}$) | 75% |\n", "| Times income ($q$) | 3 |\n", "| Maximum credit line ($Cl_{max}$) | 25,000|" ] }, { "cell_type": "code", "execution_count": 12, "metadata": { "collapsed": false, "slideshow": { "slide_type": "fragment" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[[ 1023.73054104 18750. 0. 0. ]\n", " [ 717.25781516 6749.25 0. 0. ]\n", " [ 866.65393177 12599.25 0. 0. ]]\n" ] } ], "source": [ "# The cost matrix is already calculated for the dataset\n", "# cost_mat[C_FP,C_FN,C_TP,C_TN]\n", "print data.cost_mat[[10, 17, 50]]" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Financial savings\n", "\n", "The financial cost of using a classifier $f$ on $\\mathcal{S}$ is calculated by\n", " \n", " $$ Cost(f(\\mathcal{S})) = \\sum_{i=1}^N y_i(1-c_i)C_{FN_i} + (1-y_i)c_i C_{FP_i}.$$\n", "\n", "Then the financial savings are defined as the cost of the algorithm versus the cost of using no algorithm at all.\n", "\n", " $$ Savings(f(\\mathcal{S})) = \\frac{ Cost_l(\\mathcal{S}) - Cost(f(\\mathcal{S}))} {Cost_l(\\mathcal{S})},$$\n", "\n", "where $Cost_l(\\mathcal{S})$ is the cost of the costless class" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "#Models Savings" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "## costcla.metrics.savings_score(y_true, y_pred, cost_mat)" ] }, { "cell_type": "code", "execution_count": 13, "metadata": { "collapsed": false, "slideshow": { "slide_type": "fragment" } }, "outputs": [], "source": [ "# Calculation of the cost and savings\n", "from costcla.metrics import savings_score \n", "\n", "# Evaluate the savings for each model\n", "results[\"Savings\"] = np.zeros(results.shape[0])\n", "for model in classifiers.keys():\n", " results[\"Savings\"].loc[model] = savings_score(y_test, classifiers[model][\"c\"], cost_mat_test)" ] }, { "cell_type": "code", "execution_count": 13, "metadata": { "collapsed": false, "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "# Plot the results\n", "colors = sns.color_palette()\n", "\n", "def fig2():\n", " fig, ax = plt.subplots()\n", " l = ax.plot(range(results.shape[0]), results[\"F1Score\"], \"-o\", label='F1Score', color=colors[2], linewidth=7, markersize=15)\n", " b = ax.bar(np.arange(results.shape[0])-0.3, results['Savings'], 0.6, label='Savings', color=colors[0])\n", " plt.legend(loc='center left', bbox_to_anchor=(1, 0.5), fontsize=22)\n", " ax.set_xlim([-0.5, results.shape[0]-1+.5])\n", " ax.set_xticks(range(results.shape[0]))\n", " ax.set_xticklabels(results.index)\n", " plt.tick_params(labelsize=22)\n", " plt.show()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "#Models Savings" ] }, { "cell_type": "code", "execution_count": 14, "metadata": { "collapsed": false, "scrolled": true }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4YAAAHzCAYAAAB8AwNWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig2()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "- There are significant differences in the results when evaluating a model using a traditional cost-insensitive measures" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- ~17% of savings is very bad!" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Train models that take into account the different financial costs" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "

Example-Dependent Cost-Sensitive Classification

" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "## *Why \"Example-Dependent\"" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "Cost-sensitive classification ussualy refers to class-dependent costs, where the cost dependends on the class but is assumed constant accross examples." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "In credit scoring, different customers have different credit lines, which implies that the costs are not constant" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Bayes Minimum Risk (BMR)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "The BMR classifier is a decision model based on quantifying tradeoffs between various decisions using probabilities and the costs that accompany such decisions. " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "In particular:\n", "\n", "$$ R(c_i=0|\\mathbf{x}_i)=C_{TN_i}(1-\\hat p_i)+C_{FN_i} \\cdot \\hat p_i, $$\n", "and\n", "$$ R(c_i=1|\\mathbf{x}_i)=C_{TP_i} \\cdot \\hat p_i + C_{FP_i}(1- \\hat p_i), $$" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# BMR Code" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "```\n", "costcla.models.BayesMinimumRiskClassifier(calibration=True)\n", "```" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "```\n", "fit(y_true_cal=None, y_prob_cal=None)\n", "```\n", "- Parameters\n", " - **y_true_cal** : True class\n", " - **y_prob_cal** : Predicted probabilities" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "```\n", "predict(y_prob,cost_mat)\n", "```\n", "- Parameters\n", " - **y_prob** : Predicted probabilities\n", " - **cost_mat** : Cost matrix of the classification problem. \n", "\n", "- Returns\n", " - **y_pred** : Predicted class" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# BMR Code" ] }, { "cell_type": "code", "execution_count": 15, "metadata": { "collapsed": false, "scrolled": true, "slideshow": { "slide_type": "fragment" } }, "outputs": [], "source": [ "from costcla.models import BayesMinimumRiskClassifier\n", "ci_models = classifiers.keys()\n", "\n", "for model in ci_models:\n", " classifiers[model+\"-BMR\"] = {\"f\": BayesMinimumRiskClassifier()}\n", " # Fit\n", " classifiers[model+\"-BMR\"][\"f\"].fit(y_test, classifiers[model][\"p\"])\n", " # Calibration must be made in a validation set\n", " # Predict\n", " classifiers[model+\"-BMR\"][\"c\"] = classifiers[model+\"-BMR\"][\"f\"].predict(classifiers[model][\"p\"], cost_mat_test)" ] }, { "cell_type": "code", "execution_count": 15, "metadata": { "collapsed": false, "scrolled": true, "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "for model in ci_models:\n", " # Evaluate\n", " results.loc[model+\"-BMR\"] = 0\n", " results.loc[model+\"-BMR\", measures.keys()] = \\\n", " [measures[measure](y_test, classifiers[model+\"-BMR\"][\"c\"]) for measure in measures.keys()]\n", " results[\"Savings\"].loc[model+\"-BMR\"] = savings_score(y_test, classifiers[model+\"-BMR\"][\"c\"], cost_mat_test) " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# BMR Results" ] }, { "cell_type": "code", "execution_count": 16, "metadata": { "collapsed": false, "slideshow": { "slide_type": "-" } }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4YAAAHzCAYAAAB8AwNWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig2()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# BMR Results\n", "\n", "- Bayes Minimum Risk increases the savings by using a cost-insensitive method and then introducing the costs\n", "- Why not introduce the costs during the estimation of the methods?" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "## Cost-Sensitive Decision Trees (CSDT)\n", "\n", "A a new cost-based impurity measure taking into account the costs when all the examples in a leaf\n", "\n", "```\n", "costcla.models.CostSensitiveDecisionTreeClassifier(criterion='direct_cost', criterion_weight=False, pruned=True)\n", "```" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "## Cost-Sensitive Random Patches (CSRP)\n", "\n", "Ensemble of CSDT\n", "\n", "```\n", "costcla.models.CostSensitiveRandomPatchesClassifier(n_estimators=10, max_samples=0.5, max_features=0.5,combination='majority_voting)\n", "```" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# CSDT & CSRP Code" ] }, { "cell_type": "code", "execution_count": 24, "metadata": { "collapsed": false, "slideshow": { "slide_type": "fragment" } }, "outputs": [], "source": [ "from costcla.models import CostSensitiveDecisionTreeClassifier\n", "from costcla.models import CostSensitiveRandomPatchesClassifier\n", "\n", "classifiers = {\"CSDT\": {\"f\": CostSensitiveDecisionTreeClassifier()},\n", " \"CSRP\": {\"f\": CostSensitiveRandomPatchesClassifier()}}\n", "\n", "# Fit the classifiers using the training dataset\n", "for model in classifiers.keys():\n", " classifiers[model][\"f\"].fit(X_train, y_train, cost_mat_train)\n", " classifiers[model][\"c\"] = classifiers[model][\"f\"].predict(X_test)" ] }, { "cell_type": "code", "execution_count": 24, "metadata": { "collapsed": false, "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "for model in classifiers.keys():\n", " # Evaluate\n", " results.loc[model] = 0\n", " results.loc[model, measures.keys()] = \\\n", " [measures[measure](y_test, classifiers[model][\"c\"]) for measure in measures.keys()]\n", " results[\"Savings\"].loc[model] = savings_score(y_test, classifiers[model][\"c\"], cost_mat_test)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# CSDT & CSRP Results" ] }, { "cell_type": "code", "execution_count": 25, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3oAAAHzCAYAAAB2c29BAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig2()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# Lessons Learned (so far ...)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Selecting models based on traditional statistics does not give the best results in terms of cost" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Models should be evaluated taking into account real financial costs of the application" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "- Algorithms should be developed to incorporate those financial costs" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "
\n", "\n", "
\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# CostCla Library\n", "\n", "- **CostCla** is a Python open source cost-sensitive classification library built on top of Scikit-learn, Pandas and Numpy. \n", "\n", "- Source code, binaries and documentation are distributed under 3-Clause BSD license in the website http://albahnsen.com/CostSensitiveClassification/" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# CostCla Algorithms\n", "\n", "- Cost-proportionate over-sampling [Elkan, 2001] \n", "\n", "- SMOTE [Chawla et al., 2002] \n", "\n", "- Cost-proportionate rejection-sampling [Zadrozny et al., 2003]\n", "\n", "- Thresholding optimization [Sheng and Ling, 2006] \n", "\n", "- Bayes minimum risk [Correa Bahnsen et al., 2014a] \n", "\n", "- Cost-sensitive logistic regression [Correa Bahnsen et al., 2014b] \n", "\n", "- Cost-sensitive decision trees [Correa Bahnsen et al., 2015a] \n", "\n", "- Cost-sensitive ensemble methods: cost-sensitive bagging, cost-sensitive pasting, cost-sensitive random forest and cost-sensitive random patches [Correa Bahnsen et al., 2015c]" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "# CostCla Databases\n", "\n", "- Credit Scoring1 - Kaggle credit competition [Data], cost matrix: [Correa Bahnsen et al., 2014] \n", "\n", "- Credit Scoring 2 - PAKDD2009 Credit [Data], cost matrix: [Correa Bahnsen et al., 2014a] \n", "\n", "- Direct Marketing - PAKDD2009 Credit [Data], cost matrix: [Correa Bahnsen et al., 2014b] \n", "\n", "- Churn Modeling, June 2015 " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "#Future Work\n", "\n", "- CSDT in Cython\n", "- Cost-sensitive class-dependent algorithms\n", "- Sampling algorithms\n", "- Probability calibration (Only ROCCH)\n", "- Compatibility with Python $\\ge$ 3.4\n", "- Other algorithms\n", "- More databases" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "You find the presentation and the IPython Notebook here:\n", " \n", "* http://nbviewer.ipython.org/format/slides/github/\n", "albahnsen/CostSensitiveClassification/blob/\n", "master/doc/tutorials/slides_edcs_credit_scoring.ipynb#/\n", "* https://github.com/albahnsen/CostSensitiveClassification/ blob/master/doc/tutorials/slides_edcs_credit_scoring.ipynb\n", "\n", "This slides are a short version of this tutorial:\n", "\n", "* http://nbviewer.ipython.org/github/albahnsen/CostSensitiveClassification/ blob/master/doc/tutorials/tutorial_edcs_credit_scoring.ipynb" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "

Thanks!

\n", "
\n", "\n", " \n", " \n", "\n", "
\n", "\n", " \n", " \n", " \n", " \n", "\n", "al.bahnsen@gmail.com \n", "
\n", " \n", "\n", " \n", " \n", " \n", "\n", "\n", "\n", "http://github.com/albahnsen \n", "\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "http://linkedin.com/in/albahnsen \n", "\n", "
\n", "\n", "\n", " \n", " \n", " \n", "\n", " \n", "@albahnsen \n", "\n", "
\n", "
" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "collapsed": false, "slideshow": { "slide_type": "skip" } }, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "#Format from https://github.com/ellisonbg/talk-2013-scipy\n", "from IPython.display import display, HTML\n", "s = \"\"\"\n", "\n", "\n", "\"\"\"\n", "display(HTML(s))" ] } ], "metadata": { "celltoolbar": "Slideshow", "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": "3.4.3" } }, "nbformat": 4, "nbformat_minor": 0 }