1. Jennifer Lewis Priestley Presentation of “Assessment of Evaluation Methods for Prediction and Classification of Consumer Risk in the Credit Industry” co-authored with S. Nargundkar. (Accepted for publication as a chapter in "Neural Networks for Business Forecasting" by Peter Zhang, PhD (Ed))

2. Objectives This paper addresses the answers to two important questions: 1.Does model development technique improve classification accuracy? 2.How will model selection vary based upon the evaluation method used?

3. Objectives Discussion of Modeling Techniques Discussion of Model Evaluation Methods Empirical Example

4. Model Development Techniques Modeling plays an increasingly important role in CRM strategies: Target Marketing Response Models Risk Models Customer Behavioral Models Usage Models Attrition Models Activation Models Collections Recovery Models Product Planning Customer Acquisitio n Customer Acquisitio n Customer Management Customer Management Creating Value Creating Value Collection s/Recover y Other Models Segmentation Models Bankruptcy Models Fraud Models

5. Model Development Techniques Given that even minimal improvements in model classification accuracy can translate into significant savings or incremental revenue, many different modeling techniques are used in practice: Statistical Techniques Linear Discriminant Analysis Logistic Analysis Multiple Regression Analysis Non-Statistical Techniques Neural Networks Cluster Analysis Decision Trees

6. Model Evaluation Methods But, developing the model is really only half the problem. How do you then determine which model is best?

7. Model Evaluation Methods In the context of binary classification (one of the most common objectives in CRM modeling), one of four outcomes is possible: 1. True positive (a “good” credit risk is identified as “good”) 2. False positive (a “bad” credit risk is identified as “good”) 3. True negative (a “bad” credit risk is identified as “bad”) 4. False negative (a “good” credit risk is identified as “bad”)

8. Model Evaluation Methods If all of these outcomes, specifically the errors, have the same associated costs, then a simple global classification rate is a highly appropriate evaluation method: 65050700 200100300 8501501000 True GoodTrue BadTotal Predicted Good Predicted Bad Total Classification Rate = 75% ((100+650)/1000)

9. Model Evaluation Methods The global classification method is the most commonly used, but fails when the costs of the misclassification errors are different (Type 1 vs Type 2 errors) - Model 1 results: Global Classification Rate = 75% False Positive Rate = 5% False Negative Rate = 20% Model 2 results: Global Classification Rate = 80% False Positive Rate = 15% False Negative Rate = 5% What if the cost of a false positive was great, and the cost of a false negative was negligible? What if it was the other way around?

10. Model Evaluation Methods If the misclassification error costs are understood with some certainty, a cost function could be used to evaluate the best model: Loss=π 0 f 0 c 0 +π 1 f 1 c 1 Where, π i is the probability that an element comes from class i, (prior probability), f i is the probability that an element will be misclassified in i class, and c i is the cost associated with that misclassification error.

11. Model Evaluation Methods An evaluation model that uses the same conceptual foundation as the global classification rate is the Kolmorgorov-Smirnov Test:

12. Model Evaluation Methods What if you don’t have ANY information regarding misclassification error costs…or…the costs are in the eye of the beholder?

13. Model Evaluation Methods The area under the ROC (Receiver Operating Characteristics) Curve is an option: 1-Specificity (False Positives) Sensitivity (True Positives) 0 1 1 θ=.5 θ=1.5<θ<1

14. Empirical Example So, given this background, the guiding questions of our research were – 1. Does model development technique impact prediction accuracy? 2. How will model selection vary with the evaluation method used?

15. Empirical Example We elected to evaluate these questions using a large data set from a pool of car loan applicants. The data set included: 14,042 US applicants for car loans between June 1, 1998 and June 30, 1999. Of these applicants, 9442 were considered to have been “good” and 4600 were considered to be “bad” as of December 31, 1999. 65 variables, split into two groups – Transaction variables (miles on the vehicle, selling price, age of vehicle, etc.) Applicant variables (bankruptcies, balances on other loans, number of revolving trades, etc.)

16. Empirical Example The LDA and Logistic models were developed using SAS 8.2, while the Neural Network models were developed using Backpack® 4.0. Because there is no accepted guidelines for the number of hidden nodes in Neural Network development, we tested a range of hidden layers from 5 to 50.

17. Empirical Example Quick Review on Linear Discriminant Analysis: General Form: Y=X 1 + X 2 + X 3 …+X n  The dependent variable (Y) is categorical (can be 2 or more categories)…the independent variables (X) are metric;  The linear variate maximizes the discrimination between two pre-defined groups;  The primary assumptions include: Normality Linearity Non-multicollinearity among the independent variables  The discriminant weights indicate the contribution of each variable;  Traditionally a “hit” matrix is the output.

18. Empirical Example Quick Review on Logistic Analysis: General Form: Prob event /Prob non-event = e B0+B1X1+B2X2…+BnXn  The technique requires a binary dependent variable;  Is less sensitive to assumptions of normality;  Function is S-shaped and is bounded between 1 and 0;  Where LDA and Regression use the least squares method of estimation, Logistic Analysis uses a maximum likelihood estimation algorithm;  The weights are measures of changes in the ratio of the probabilities or odds ratios;  Proc Logistic in SAS produces a “classification” matrix that provides sensitivity and specificity information to support the development of an ROC curve.

19. Empirical Example Quick Review on Neural Networks: Input Layer Hidden Layer Output Layer Σ S Combination Function combines all inputs into a single value, usually as a weighted summation Transfer Function Calculates the output value from the combination function input output

20. Empirical Example - Results TechniqueClass Rate “Goods” Class Rate “Bads” Class Rate “Global” ThetaK-S Test LDA 73.91% 43.40% 59.74% 68.98% 19% Logistic70.54%59.64% 69.45% 68.00%24% NN-5 Hidden Layers63.50%56.50%58.88%63.59% 38% NN-10 Hidden Layers75.40%44.50%55.07%64.46%11% NN-15 Hidden Layers60.10%62.10%61.40%65.89%24% NN-20 Hidden Layers62.70%59.00%60.29%65.27%24% NN-25 Hidden Layers 76.60% 41.90%53.78%63.55%16% NN-30 Hidden Layers52.70% 68.50% 63.13%65.74%22% NN-35 Hidden Layers60.30%59.00%59.46%63.30%22% NN-40 Hidden Layers62.40%58.30%59.71%64.47%17% NN-45 Hidden Layers54.10%65.20%61.40%64.50%31% NN-50 Hidden Layers53.20% 68.50% 63.27%65.15%37%

21. Empirical Example - Conclusions What were we able to demonstrate? 1.The “best” model depends upon the evaluation method selected; 2.The appropriate evaluation method depends upon situational and data context; 3.No multivariate technique is “best” under all circumstances.