Using self-organizing maps for analyzing credit rating and financial ratio data

Credit rating granting in rating agencies is a complex decision making process for outside rating information users. The data these rating agencies required for making rating decision cover many facets, including financial data, operations data, the data regarding interview with top managers of issuers, etc. The systematic rating process, such as how each data item contributes to each specific rating granting, is still blind to the outsiders. Therefore, this study proposes an easy analytical tool for visualizing the relationships between credit rating information and financial ratio data. Self-organizing maps (SOMs) have been effectively used for visualizing and clustering tasks in numerous applications, such as financial statement analysis and document analysis, and thus this study applies SOMs on analyzing the relationship patterns. Banking industry data are used as the test bed. The study results demonstrate that the SOM could be a feasible tool for uncovering the relationships between those rating symbols and the data referred by rating agencies.