Analysis of Borrower Risk Characteristics Using Canopy-K-means Clustering

Technological advancements, shifts in social trends, and deregulation in the financial sector have led to a rapid growth in the fintech industry. For example, fintech platforms that offer loan rate comparison services attract customers by providing opportunities to compare and choose among various loan options. As the market for online personal loan comparison services expands, it becomes more important for lending institutions to understand and manage fintech service customers. In this study, we analyzed the characteristics of borrowers who used loan comparison services by performing clustering analysis. We employed K-means clustering with canopy pre-training algorithm as a pre-processing step and identified nine distinct customer clusters. These clusters were categorized into three groups (good, risky, and high-risk) based on customer age and average loan workout ratios. The characteristics of each cluster were analyzed and presented. The outcomes of this study are expected to contribute to the enhancement of risk management and marketing processes for financial institutions.