Loan Credit Risk System
Loan Credit Risk System
OVERVIEW
- Hawaiian Telcom is the primary provider of comprehensive communications and entertainment services, solutions, and products in Hawaii. With its headquarters based in Honolulu, Hawaiian Telcom is the prominent integrated communications provider for both residential and business customers in the region.
- Hawaiian Telcom collaborated with Microsoft - OCP and People Tech to develop machine learning model for assessing loan credit risk to determine whether to grant or decline loan requests.
- The company figured out the need for:
- Need for accurate credit risk assessment: Hawaiian Telcom faced the challenge of effectively evaluating loan requesters and assessing the risk of default to make informed decisions on loan approvals.
- Data integration and analysis: The integration of borrower's financial history and loan information required a robust solution to derive insights for risk assessment.
- Deployment infrastructure: The model was deployed in a hybrid architecture, leveraging Azure Cloud and Azure Stack to ensure scalability and efficiency.
SOLUTION PROVIDED BY PEOPLE TECH GROUP
- Imputation of missing values: Initial imputation techniques were applied to handle missing values in the merged data.
- Conditional Inference Trees: Optimal cuts were defined for categorical variables using conditional inference trees, enabling accurate credit risk assessment.
- Machine Learning Model: Standard logistic regression was used as the machine learning model, with AUC (Area Under the Curve) serving as the final evaluation metric.
BENEFITS
- Streamlined loan approval process: The implementation of the machine learning solution automated the claims process for approximately 82% of Hawaiian Telcom's customers, improving operational efficiency.
- Enhanced risk assessment: Accurate credit risk assessment allowed Hawaiian Telcom to make informed decisions on loan approvals, mitigating the risk of default and improving overall portfolio performance.
