crspl.info@careedge.in
Category:
Consulting
Date:
10 March 2023
Implementing Predictive Credit Risk Analytics

Client Profile

Client is the largest private sector bank in Sri Lanka and known as the benchmark private sector bank in the country with over 100 years of unparalleled growth & achievement. This case study showcases how our platform used the Machine Learning techniques to Analyse & Predict the borrower’s Credit Risk Rating for the client bank.

Objective

To analyze and determine risk levels involved on credits, finances and loans.
To assist the analysts in predictive risk analysis.
Increase effectiveness of risk management through measuring the riskiness of borrower.

Problem Statements

Collecting historical financial & non-financial data from various internal & external sources.
Slicing & dicing for analyzing data in different views and perspectives.
CRSPL ensured that the Risk Management policy and processes are appropriately and adequately aligned to the client’s future growth plans
Use of efficient feature extraction technique to dimension reduction.
Perform model validation / cross validation to get better & accurate credit rating.

Solutions

Effective data ingestion techniques to collect data from various internal & external sources.
Better data visualization tools for correlation analysis, feature extraction and data aggregation.
Used models from probabilistic classifiers and ensembled learning to create classification algorithms.
Used various metrics to interpret the predictive validity of a model (e.g., R-square, mean squared error, sensitivity, goodness of fit, ROC, loss function,  confusion matrix), and validated the models by using methods such as, Cross-validation and bootstrap.
Periodic review of the model and train the model based on the latest data.

Benefits

With the proposed solution, client was able to calculate the credit rating for its borrowers with accuracy of around 88%.
It helped the client to improve the credit monitoring process by predicting the customer’s delinquency and accordingly categorize them into ‘good account’ (one who pay on time) and ‘bad account’ (one who default) with quick turnaround time.
Implementation in Bhutan – FTP
Case Study
Next
Case Study’s

Our Similar Consulting Case Study’s

Finance, Tax

Financial Reporting
Implementation in Bhutan – FTP
View details

Business, Finance

Financial Reporting
Implementing Predictive Credit Risk Analytics
View details

Business, Finance, Tax

Financial Reporting
Risk Management Consulting Framework
View details