Research On Stock Value Investment Based On Artificial Intelligence

Value investment theory has been the development history of several decades since its inception date and has become the mainstream investment philosophy in the western mature market now. With the increasing development of China's stock market, the trend of the market returning to ration being obvious. Under such market environment, value investment is being pursued by China stock market investors gradually. Meanwhile the differences in the domestic and foreign stock markets and the qualities of the listed companies lead much criticism to the imported value investment theory. However, with China's securities market being standard gradually, the value investment-oriented concept must become the mainstream investment philosophy. In such background, to deepen the study of the value investment theory has theoretical and practical significance from the reality of China's securities market. From the theoretical point of view, the study is benefit to improve the maturity and perfection of the value investment philosophy and to cultivate the value investment philosophy suited to China's stock market. Based on the understanding of the stock market, the work puts up in two aspects of theory and method.From the basic opinion of the traditional value investment, the investment decision of value investment is based on "safety marginal" between the inherence value and the stock price, especially pay attention to the value of the stock, the judgment of the stock inherence value always obey the investment theory of Williams that is "the inherence value of stocks equal to the present value of the dividend and interest in the future years", this theory measure the inherence value by the present value of stock return but lack of maneuverability, there is another reasonable research incline about value investment which is judge the stock "safety marginal" by price-earnings ratio and price-net assets ratio, but the index of P/E and P/B can't reflect the value and growth of the stock, the ignoring of stock growth is also the great weakness of traditional value investment theory. Against the weakness of traditional value investment theory, this paper propose the concept of stock virtual value from the consideration of stock virtuality and analyze the changeability and stability, then this paper propose the new value investment frame as the supplement and development of the traditional value investment theory. Compare with the traditional value investment theory the new value investment theory is based on the stock virtual value, the listed firm's operation condition affect the virtual value of the stock directly by the market effect, moreover affect the extent of stock's changeability and stability, thereby, new value investment theory master the relationship and developing trend between stock virtual value and price by the analyzing of listed firm's financial reports. The purpose of new value investment is digging out the stocks which the stock price under the virtual value and regressed to it, this is the strategy of stock investment. The new value investment considers the changeability and stability at the same time with strong manipulation especially used to the stock market of China, it is a kind of important complementarities for the traditional value investment theory.Under the new value investment frame, this paper combine the pattern recognize and new value investment and design the method of value investment during which the artificial intelligence impenetrate the paper. Because in the process of stock investment, the investors can't get all the information, under the new value investment method this paper take the advantage of fuzzy clustering to deal with the great volume datum. Dig out the key information from the listed firm's financial reports, from the empirical aspects, propose the simple stock characteristic sets for the support vector machine stock value investment choosing model, the characteristic sets basically contain all the available information of stock value investment. Based on this, the paper constructs the support vector machine stock value investment choosing model by empirical method which can choose the stock effectively by the judgment of valuable information. For the empirical research, this paper recognize the stocks of SSE-180 index by support vector machine stock value investment choosing model, and compare the return of half year hold, one year hold, two and three years hold to the SSE-180 index return, the result shows that more than 70 percent stocks' return are high than index return, the result confirm the stock choosing capacity of support vector machine stock value investment choosing model. Although the support vector machine stock value investment choosing model could choosing the potential stock in a high probability, but the results may also include some failed stocks, it is important to construct the portfolio to dissemination the investment risk. This paper proposes the RAROC target optimize investment model under the investment portfolio mean-variance model of Markowitz and solve the investment portfolio by simulated annealing genetic algorithm. This model use the VAR as the risk measurement and make the RAROC as the target optimize function which could conformity the dual target restriction of investment decision and return as a RAROC single target restriction which could disseminate the signal stock centralize risk effectively.