Predicting And Managing Credit Risk By Implementing Scorecard Using Hybrid Strategy With Trust Rating

摘要

Financial institutions possess a great deal of credit risk in assessing credit application for approval. In recent days, to assess, manage and to make decisions on the credit risks of the customers, financial institutions employ internal scorecards. However, major banks use several existing one-dimensional credit scoring model which may lead to inaccurate assessment results. In this paper, a three-dimensional hybrid credit scoring technique has been proposed that includes sequential application scoring along with the dual credit scoring matrix model. Dual credit scoring model uses behavioural credit scoring and credit bureau scoring for computing the trust rating. Also, the behavioural scoring model employs an optimized multiple rank score based feature selection for accurate scoring. On employing the signed approach based trust ratings, the customers are categorized into three risk groups for assessing and managing the customer credit risks. The credit strategies to be followed in making decisions are also presented along with the empirical analysis. The results from the analysis show that the proposed method provides 88% precision with 43.17 K-S statistics value.