PREDICTIVE ANALYTICS FRAMEWORK FOR DYNAMIC CREDIT LINE OPTIMIZATION USING BEHAVIORAL DATA

Abstract

The increasing complexity of consumer financial behavior has created a need for more intelligent and adaptive credit management systems. Traditional credit line assignment methods primarily rely on static credit scores and historical financial records, which may not adequately capture changes in customer behavior and risk profiles. This study proposes a Predictive Analytics Framework for Dynamic Credit Line Optimization Using Behavioral Data to improve credit decision-making processes in financial institutions. The framework utilizes customer behavioral indicators such as credit utilization ratio, repayment history, transaction frequency, spending patterns, and account activity to predict creditworthiness and optimize credit limits dynamically. A quantitative research approach was adopted, involving the analysis of behavioral and financial data collected from a hypothetical sample of 500 customers. Various machine learning algorithms, including Logistic Regression, Decision Trees, Random Forest, Gradient Boosting, and Artificial Neural Networks, were evaluated for predictive performance. The results revealed that behavioral data significantly enhances the accuracy of credit risk prediction and credit line optimization. Among the tested models, Random Forest and Gradient Boosting demonstrated superior predictive capabilities. The findings indicate that dynamic credit line optimization can improve risk management, reduce default probabilities, enhance customer satisfaction, and increase profitability for financial institutions.