Social Credit Scoring System for a personal loan provider
India’s first mobile app based personal loan provider, uses a Machine Learning based credit scoring model for assessing credit worthiness of a customer. The model combines traditional financial data with social data to change the lending market in India.
Business Impact
Automated 80% of the loan application decisions.
Processed loan applications within seconds as opposed to hours.
Identified the fraudulent applicants in real time and prevented significant amount of fraud.
Methodology
A credit scoring model that combines financial as well as social data of the applicant. The model uses social data from Facebook, phone contacts, text messages, GPS location, call records; and financial data from credit bureau and bank statements. Furthermore, fraudulent applications are detected by analyzing social and financial behavior. The model performs better than human expert underwriters in identifying people who are likely to default on the loan.