Research And Design Of Arbitrage Stock Selection Model Based On Data Mining

The Chinese securities market has been undergoing a rapid development in recent ten years with China’s transformation from planned economy to a market economy. By far, Shanghai and Shenzhen Stock Exchange has owned more than one thousand and six hundred quoted companies The market will be more developed with the emergence of stock index futures, ETF and other financial derivatives as well as the arrival of all-round circulation age. It will be more and more necessary to introduce a variety of new advanced, innovative, quantitative research methods into the market. At the same time, a list of new trading methods existed in foreign markets such as hedging and arbitrage have gradually run into the eyes of domestic investors accompanied by the appearance of the short mechanism. However, data models designed for such transactions are rare.As a newly developed technique, the data mining method have been used successfully in lots of areas including marketing, sales, and customer resource management. But there is still not enough study to apply the data mining method to stock selection model building, portfolio profit forecast and investment model building.The essay introduces data mining method into the stock picking investment field. Mainly discussed the use of decision tree algorithm to the listed company package stock alpha value of data mining. Draw the short term alpha value run win a package of stock market, use and futures or margin hedge formation arbitrage model. In establishing the model after, it also adopted the "data resolution" to determine the effect of these predictions steady and reliable.Incorporated with widely used statistics and econometrics mcthods, the data mining method could reach better quantitative effort and enable the investors to obtain a stable return on investment in the more mature market.