Research On Robust Stock Value Investing Based On Ordered Mechanism

The global financial crisis in2007derived from excessive expansion of financial derivative products causes people to re-examine the stock in-vestment decisions. Is the effective choice for the underlying assets more important? Can stock selection strategy beat the market? As the stock market is becoming increasingly important but more risky as well, these issues has become the major focus of academics and practitioners, and de-cision making of stock selection has naturally become a key basic problem of stock investment decisions. Among the large number of stock investing strategies, value investing attracts most attention. Based on the clas-sic value investing theory, the stocks are generally categorized as value stocks and growth stocks in the practice of real investment. The exist-ing researches has provided us with a lot of evidence suggesting that value stocks have value premium effect which can generate excess returns. How-ever, the debate on the value premium effect has never stopped. There is evidence showing that value premium effect based on the classical value characteristics is instable, or that value stocks are the companies with risk factors. It can be seen all the controversy shows that value stock invest-ment model based on the classic value investing theory does not have the characteristics of being robust. Obviously, in the background of the reces-sion of global real economy and exacerbated capital market risk, study on robust stock value investing model has important theoretical significance and practical value.For decision-makers, the stock performance are their major concern when selecting stocks. However, in real stock market, the factors affecting the stock performance are too numerous to mention. So, what kind of factors is the most essential? And what kind of mode is the most robust? This dissertation holds that the enterprises’economic operation efficiency, as basis of business value, is the essential factor of stock performance, and stocks excellent in both operation and stock market are the most "reliable" and "robust" ones. Taking the features of being "robust" into consideration, this disser-tation chooses the accounting information:a comprehensive measure of companies’performance as basis, robust features:excellent performance of operating and stock as the core, and ordered solving:robust stock value investing model as context to establish the theoretical framework of robust stock value investing. Essentially, robust stock value investing is a typical problem of multiple attribute ranking decision making. To ensure a more comprehensive and effective portrait of the data characteristics, this dissertation uses interval data as the basic representation of data. In order to effectively establish the ordered solving model, this disserta-tion focuses on the three characteristics-robust, localized and overall-to establish the ordered mechanism of multiple attribute ranking decision making; for the purpose of effectively carrying out the ordered solving, it is crucial to address the following core issues:●How to build effective complete rank method of interval data?●How to establish scientific method of feature evaluation?●How to select the key features of interval ordered decision table?On the basis of the theoretical framework of robust stock value in-vesting, focusing on the three core scientific issues mentioned above, this dissertation uses interval ordered information system as description frame-work of decisions problems of robust stock value investing under the frame-work of rough set theory, and attains methods of robust stock value invest-ing ranking decision. The main results and innovations can be summarized as follows:1. Based on five aspects of accounting information including prof-itability, cash flow capacity, operating capacity, development capacity and debt-paying ability, a company’s operation performance can be measured comprehensively. Based on the feature of excellent both in operation and stock, robust stock value investing model can be built, which further en-riches and develops the theory of value investing.2. Based on the dominance relation under the framework of rough set theory, this dissertation considers human cognitive model of local order-ing, and introduces superiority degree ranking as priority criterion. Based on the overall ordering characteristic, and focusing on a more precise por-tray of preferability degree among objects, this dissertation introduces the directional distance index, and thereby establishes the complete rank methods of interval data’s two-grade ranking decision.3. Facing the interval ordered information system, and considering the complementary set feature of information gain, this dissertation pro-poses concepts of complementary entropy, complementary mutual infor-mation and complementary conditional entropy of interval ordered, estab-lishes the entropy measure system of uncertainty representation of interval ordered information system, and provides an effective feature selection and feature evaluation method for robust stock value investing ranking decision.Overall, based on the ordered mechanism, this dissertation carries out the research for robust stock value investing decision from theoretical and technical perspectives. The empirical study on Shanghai stock exchange180index components suggests that:the robust stock value investing model can obtain a stable excess return, which verifies its validity and robustness of robust stock value investing ranking decision. More gener-ally, the integration of feature selection, feature evaluation, and modeling method of complete rank provides a robust complete rank for risk-averse decision-makers, and further enriches the theories and methods of artifi-cial intelligence in decision-making.