A Decision Support Model for Online Lending Creditworthiness Using Comparative Personality Indicators

The rise of online lending platforms has improved financial accessibility but also increased credit default risk due to information asymmetry and limited borrower profiling. Traditional creditworthiness models rely primarily on financial and demographic data, which often fail to capture behavioral characteristics. This study proposes a decision support model for creditworthiness prediction by integrating personality indicators from the Big Five Personality Traits and the California Psychological Inventory (CPI). The framework incorporates these personality-based features into a machine learning-driven system alongside traditional borrower data. Psychological indicators are quantified and assessed using multiple classification models to evaluate their impact on predictive performance. The model's effectiveness is measured using metrics such as accuracy, precision, recall, F1-score, and Area Under the Curve (AUC). Empirical results show a significant improvement in prediction accuracy, with the AUC rising from 0.74 in the baseline model to 0.87 after including personality features. A comparative analysis reveals the relative contributions of each personality framework, demonstrating that personality indicators enhance predictive performance over traditional models. These findings emphasize the value of incorporating behavioral factors, supporting the development of more effective and sustainable credit risk assessment systems.