 |
An International Meeting of American Accounting
Association
|
An Analysis Of The “Going Concern Assumption”: Text Mining From Japanese Financial Report |
Cindy Yoshiko Shirata
Shibaura Institute of Technology
Manabu Sakagami
Osaka City University
Abstract: Signs of a corporate failure are bound to manifest themselves not only in financial data but also in non-financial data. This study detected key words that can signal the presence of “questions about a corporation’s ability to continue as a going concern” by analyzing non-financial data, qualitative data, that are included in public financial report (Japan 10-K) released by the corporation. The analysis was made by applying text-mining techniques that have been gaining increasing attention in the field of artificial intelligence. The result of the analysis revealed that non-financial data that is released by continuing corporations show a unique tendency, and the key words that appear with high frequencies among continuing corporations reflect their sound financial trends. In other words, the study verified that it is possible to assess the possibility that a corporation will continue to operate, based on non-financial figures, without looking at financial figures.