earticle

영어

Current corporate credit rating methods tend to focus on rating single corporations and usually fail to assess corporate groups from different industries or geographical areas. This paper presents a method of evaluating corporate credit ratings using the community detection of label propagation (CCRELP). First, the study used a label propagation algorithm to group corporations with similar characteristics into communities based on the connections between the companies, their industries, and their geographical locations. Then, the study performed an evaluation using various indicators and weights to acquire credit ratings for companies from different communities. Experiments using real-world datasets showed that, when using various indicators to rate corporate credit, CCRELP could detect differences among the results for companies from different communities, thus indicating the possibility of developing a more reasonable evaluation system for companies with similar characteristics. The results also proved that the community-based credit rating system is an effective and feasible means of corporate credit rating.