Application Of Svm In The Quantitative Investment Of Stock Index Futures In Shanghai And Shenzhen Stock Index 300

A kind of new investment mode which combines the quantitative method and the modern financial theory and the investment practice and uses the computer technology to carry on the transaction.What is the essence of quantitative investment,but the quantitative model of the thought and logic of trade.The basic theoretical knowledge of quantitative investment include:multi factor,stock trend tracking,risk control,data mining,pattern recognition,market statistics,wavelet analysis,polynomial regression,time series,and calculate the volatility of capital flows,market sentiment analysis,support vector machine,market microstructure,yield distribution,and fractal theory stochastic process.In this paper,the concept of quantitative investment,the relevant content and the mainstream of the quantitative investment system are described,and a basic quantitative trading system structure and related design.And the support vector machine model,algorithm and its application in investment decision making are studied and discussed,aiming at parameter optimization of support vector machine and kernel function selection in the two key technical problems,using cross validation algorithm to optimize the parameters,and according to the Shanghai and Shenzhen 300 stock index futures market results back to historical data select the appropriate parameters,to improve the prediction performance of the model,to provide a reference for decision making.In this paper,the Shanghai and Shenzhen 300 stock index futures trading data,smoothing,that is,noise reduction.The technology used in this paper is the wavelet analysis technology,which can eliminate the noise of the transaction data and keep the effective features of the original signal.After the stock index series smoothing can remove high-frequency disturbance signal interference,making the stock index trend to better show,also makes the support vector machine learning machine learning effect is more significant and has higher classification accuracy.To determine whether to buy or sell in order to predict future trends.By using the theory of support vector machine design quantitative trading strategy model,this model includes opening rules,rules and rules of stop positions.According to the first 120 trading days after the wavelet denoising of the transaction data as a dependent variable to fit the training,after the index of the 1 trading days up and down as independent variables,so rolling training and forecasting.Back to the historical trading data from January1,2015 to December 31,2015,the annual yield of 63.91%,the maximum retracement of the results of 12.33%.But the yield volatility in the market when the transition is greater,the stability can be further enhanced.Finally,put forward the idea of improving the stability of the model and how to increase the profit and reduce the risk.