Incorporating Markov Process and Evolution Strategy on Financial Distress Rules Extraction

• Main Authors: Jui-Hsien Meng, 孟芮賢
• Other Authors: Ying-Hua Chang
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/r5xq34

碩士 === 淡江大學 === 資訊管理學系碩士班 === 102 === 　　Business environment has greatly changed since　Globalization, domestic companies faced with more impact to business operation. Corporate management is closely linked with national economics; its profitability may help on development for economics and society. On the contrary, failure or financial distress companies can largely affect banks, many other companies and people around us. It is necessary for large corporates to make early response to the financial distress and seek for reliable financial distress alert. Past studies using basic analysis on financial distress; practically, Data mining is frequently used to find the cause of financial distress. 　　This study focused on exploring financial distress alert in corporate business, and search for financial distress rules based on dynamic financial distress model. This study includes analyses on corporate financial indices utilizing Markov decision process; it could predict distress levels that corporate will faced with. Secondly, Evolution strategies (ES) has capabilities of implicit parallel search, using ES and get evaluated by Markov decision process will find financial status on every distress state. 　　It could be more efficient and more credible to find out the financial distress alert through Evolutionary computing. For corporates, financial rules level implies the severity level of financial distress. The corporate may propose different responses to solve the business operation problem; the Government may make assessment on tanking company public on listed and OTC market. Inventers may take these rules as a reference on making decision to select stock holdings, and avoid investing tank stocks.