Advanced spectral clustering for heterogeneous data in credit risk monitoring systems

TL;DR

This paper introduces Advanced Spectral Clustering (ASC), a novel method for effectively clustering heterogeneous financial and textual data in credit risk monitoring, improving cluster quality and providing actionable insights.

Contribution

We develop ASC, which integrates financial and textual similarities with an eigenvalue-silhouette optimization, advancing spectral clustering for heterogeneous data in credit risk analysis.

Findings

ASC achieves 18% higher Silhouette score than baseline.

51% of low-risk firms include 'social recruitment' in text.

ASC's robustness confirmed across multiple clustering algorithms.

Abstract

Heterogeneous data, which encompass both numerical financial variables and textual records, present substantial challenges for credit monitoring. To address this issue, we propose Advanced Spectral Clustering (ASC), a method that integrates financial and textual similarities through an optimized weight parameter and selects eigenvectors using a novel eigenvalue-silhouette optimization approach. Evaluated on a dataset comprising 1,428 small and medium-sized enterprises (SMEs), ASC achieves a Silhouette score that is 18% higher than that of a single-type data baseline method. Furthermore, the resulting clusters offer actionable insights; for instance, 51% of low-risk firms are found to include the term 'social recruitment' in their textual records. The robustness of ASC is confirmed across multiple clustering algorithms, including k-means, k-medians, and k-medoids, with {\Delta}Intra/Inter < 0.13 and {\Delta}Silhouette Coefficient < 0.02. By bridging spectral clustering theory with heterogeneous data applications, ASC enables the identification of meaningful clusters, such as recruitment-focused SMEs exhibiting a 30% lower default risk, thereby supporting more targeted and effective credit interventions.