Evidence-Supported Credit Risk Report Generation Using News-Centric Financial Knowledge Graphs

2026-07-01

AI Summary

This study addresses the limited interpretability and factual hallucinations in existing credit risk reports, which stem from the absence of explicit modeling of the relationship between news events and financial markets. To this end, we propose FinKG-News—the first company-level financial knowledge graph anchored in news events—that integrates corporate, news, and event data to serve as a reliable source of verifiable evidence. We further design a context-aware learning framework that automatically generates interpretable, fact-grounded credit risk reports across three key financial dimensions. Experimental results demonstrate that our approach improves report quality by 19%–34% over baseline methods while substantially reducing hallucinations. Human evaluations further confirm the irreplaceable value of expert judgment. The code and resources are publicly released.

Abstract

Financial markets evolve in response to real-world events reported in news, yet these drivers often remain implicit in text. To better explain market dynamics, event-market relations must be explicitly modeled through factual, company-centric, and environment-aware knowledge graphs. We present FinKG-News, a framework that automatically constructs such graphs by extracting news events as anchors linked to companies. Using FinKG-News as grounded evidence that integrates events, news, and company data, we develop an in-context learning architecture for credit risk report generation across three core financial dimensions. Automatic and human evaluations show that automated hallucination detection and quality assessment remain unreliable, making expert judgment indispensable. Our approach consistently outperforms baselines, improving quality by 19%-34% while reducing hallucinations. The source code and project resources are publicly available at: https://github.com/ichise-laboratory/FINKG-news.

Problem

• credit risk report generation
• news events
• financial knowledge graphs
• market dynamics
• hallucination reduction

Innovation

• Financial Knowledge Graph
• News-Centric Event Extraction
• Credit Risk Report Generation
• In-Context Learning
• Hallucination Reduction