CREDIT RISK EVALUATION PREDICTION SYSTEM

Authors:

1^st Dr.A. Karunamurthy, 2^nd P. Dhinakaran

1Associate Professor, Department of Master Computer Application, Sri Manakula Vinayagar Engineering College, Pondicherry 605107, India.

karunamurthy26@gmail.com

2 Professor, Department of Master Computer Application, Sri Manakula Vinayagar Engineering College, Pondicherry 605107, India.

dkdhinakaran544@gmail.com

ABSTRACT

Credit risk assessments traditionally take place after default has occurred; thus, current methods for assessing the creditworthiness of customers do not allow early intervention. This paper presents a Customer Financial Risk Prediction system by integrating supervised machine learning classification techniques into an automated real-time monitoring platform. The system identifies customers that are at risk of defaulting; uses multiple dimensions of customer financial behavior to provide a continuous risk score and classify customers as low, medium or high risk. The use of historical records validated the Random Forest classifier as yielding an overall F1 score of 0.89 and an F1 score of 0.85 for the high-risk class - significantly better than single-threshold legacy alert systems (F1 = 0.62). Role-differentiated dashboards support timely distribution of actionable insights to stakeholders (customers, risk officers, credit analysts, and administrators). Achieving accurate, timely and interpretable risk classifications using open-source technologies reduces exposure to credit defaults and promotes ethical/fair financial decision making.

Keywords: financial risk prediction, credit scoring, Random Forest, XGBoost, predictive classification, customer analytics, machine learning, fintech.