Research On Quantitative Trading Strategy Model Based On Date Mining

Nowadays,In the changing of China's capital market,as the method of China's emerging market investment,quantitative investment is drawing more and more attention.Because of the quantitative trading mode is at the initial stage in our country,if the”T + 0” transaction system is going to be restarted,it will rapidly accelerate the development of quantitative investment trading.Therefore,quantitative trading strategy model will become the core direction of our country securities market investment research.In this paper,researching on quantitative trading strategy model based on Date Mining,we select the training set stock and prediction set stock,the data includes the some stock market information,such as clinch a deal the price of 3 seconds,volume,business of direction,clinch a deal the time,and using the method of quantitative statistical model to analysis the trading strategy of the prediction set stock.In this paper,Firstly,we use the probit model to analysis the strategy of the prediction set stock and conclude that :(1)the direction of strategy and market transaction price are positively correlated,namely,it is a buy signal when the stock price is increasing,investors are preferring to purchase,when stock price is falling,it is a sell signal,investors are preferring to sell;(2)In the transaction process of the high volume,investors prefer to make a purchase transaction strategy;When the volume is low,investors prefer to sell strategy;(3)The quantitative trading strategy has strict requirements on the trend of stock price,It obtains the misjudgment rate after forecasting the prediction set.Affertewards,in order to explore the influence of different statistical model at strategy decision,the paper shows the research of the Bayesian Network Model and support vector machine(SVM)model in developping the quantitative strategy.Later,by comparing three kinds of statistical model in the application of quantitative trading strategy,we observe that the three methods of have their own advantages in statistical analysis of quantitative trading strategy.In actual analysis applications,we often use the prediction precision and explanation to illudtrate the model.We found that according to the actual condition to select different methods,applying three methods combination,maybe achieve better prediction effect.In the end of the paper,combining with the analysis of strategy model,It put forward the reasonable suggestions of perfecting the system of quantitative trading,and put forward the optimization of using Support Vector Machine(SVM)method.