10. Density
Density
Minimal increase in average interest rate from 6% to 6.8%
Would have cut losses in 2014 by $17K or 89%
11. Density
Minimal increase in average interest rate from 6% to 6.8%
Would have cut losses in 2014 by $17K or 89%
Would have cut losses from 2009 onwards by $240K or 82%
Density
12. Predictive model created from combination of
logistic regression and machine learning (SVM)
• Basic probability theory to deal with class bias
13. Predictive model created from combination of
logistic regression and machine learning (SVM)
• Basic probability theory to deal with class bias
푃 푙표푠푠 = 푃 푑푒푓푎푢푙푡 ∗ (1 − 푃 푠표푚푒푝푎푦푚푒푛푡 푑푒푓푎푢푙푡 )
14. Predictive model created from combination of
logistic regression and machine learning (SVM)
• Basic probability theory to deal with class bias
푃 푙표푠푠 = 푃 푑푒푓푎푢푙푡 ∗ (1 − 푃 푠표푚푒푝푎푦푚푒푛푡 푑푒푓푎푢푙푡 )
15. Predictive model created from combination of
logistic regression and machine learning (SVM)
• Basic probability theory to deal with class bias
• Logistic regression identified 4 features that could predict risk
• “Riskier population”
• Borrower allowed maximum interest rate
• Loan Category
• Country of applicant
16. Predictive model created from combination of
logistic regression and machine learning (SVM)
• Basic probability theory to deal with class bias
• Logistic regression identified 4 features that could predict risk
• “Riskier population”
• Borrower allowed maximum interest rate
• Loan Category
• Country of applicant
17. Higher Risk Associated with Borrowers who
entered between August 2012 and August 2013
Density
18. Higher Risk Associated with Borrowers who
entered between August 2012 and August 2013
Density
August 2012
August 2013
19. Predictive model created from combination of
logistic regression and machine learning (SVM)
• Basic probability theory to deal with class bias
• Logistic regression identified 4 features that could predict risk
• “Riskier population”
• Borrower allowed maximum interest rate
• Loan Category
• Country of applicant
• Used identified features to train
kernel SVM with 10 fold cross validation
(cut losses by 89%)
20. Conclusions
• Impact/Significance
• Project to cut losses by $48,000 over the next year
• Over 5 year period, for every $1 million invested, recovers additional
$110,000 that can continue to be reinvested
• Actions already taken
• Implement the model the risk model for interest rates
• Actions to be taken
• Find policy change that allowed for risky population
23. Next things to do
• Incorporate fraud risk
• Incorporate a risk of default (this is likely based upon sift score or
some other metric)
• Look at interacting terms (type of business/country)