Application of clustering algorithms to improve the SVM+Prototypes for analyzing the financial statement audit risk

• Main Authors: Shiu, Chauneng, 許超能
• Other Authors: Pai, Pingfeng
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/86511475302082141728

碩士 === 國立暨南國際大學 === 資訊管理學系 === 100 === The financial statements provide investors understand the business of operating results and financial position. In recent years, the financial statements of problem are frequent and affect not only the investor can’t make immediate judgment based on the financial statements but also consumption of many resources of the community. This study use financial information and transparency to judge the financial statements whether has audit risk, first using support vector machine to predict the risk of financial statement audits, then use the SVM + Prototypes for support vector machines to extract rules. Finally, using Fuzzy C-means and self-organizing map hope can be improved the SVM + Prototypes. Researching has shown that using Fuzzy C-means more effective to increase the accuracy and coverage but self-organizing map does not. A judgment in accordance with better rule to determine a company's financial statements whether has the audit risk and explain, then give a basis for judgment to the users of financial statements.