Research On Short-term Price Prediction Of Stock Index Futures Based On Support Vector Machine

Stock index futures are the futures varieties with the index of stock price as the subject matter.In April,,the first stock index futures of our country,the CSI stock index futures contract,was officially launched,and opened the prelude to the development of stock index futures in China.With the increase of investors in the stock index futures market,stock index futures trading volume is also growing,the stock index futures play a price discovery,risk aversion effect,but also enrich the investment strategy for investors,the stock index futures market has become an important part of the financial market in our country,so it has important significance for predicting stock index futures price the.For price prediction,most of the predecessors used technology analysis,basic analysis,traditional time series and so on,but these methods often have large limitations,and the prediction results are often not ideal.Support vector machine is developed from the statistical learning theory based on structural risk minimization.Support vector machine has some advantages that other prediction models do not possess to a certain extent,so it has better effect for price prediction.In this paper,the support vector machine is used to model the historical data of stock index futures in China,and the model is used to predict the price.This paper selects the historical transaction data of China Shanghai and Shenzhen stock index futures,and divides the selected data into two parts,which are the training data and the fresh data.Select the relevant indicators that affect the futures contract price,choose eight indexes related to short-term price of stock index futures: the lowest price,the highest price,the closing machine,opening price,increase,amplitude,volume and position.The selection of the position index is the innovation of this article,and the position is the specific index of the futures variety,and it has a close relation with the price.The selection of model parameters and the selection of kernel function of supportvector machines will have a great influence on the prediction results.In terms of parameter selection,we use genetic algorithm and particle swarm optimization to optimize parameters,and evaluate the goodness and inferiority of parameters by means of mean square error index.We combine two parameters and four kernel functions to build eight regression prediction models.In the training data range,using eight models,the first eight indicators as independent variables,the stock index futures will be second days of the closing price as the independent variable,the model is trained,trained the model to predict the new data,get the forecasting result,the prediction results to choose the best predictive model selection.In order to study the applicability of the model,we choose the best model to predict Shanghai Stock Index index futures,and predict the applicability of the model through the prediction results.In this paper,we find the parameter optimization based on genetic algorithm through empirical analysis.The SVM regression prediction model combined with radial basis function is superior to the other seven models in short-term price prediction of stock index futures.The proposed model also shows ideal results on other samples,which proves that the model has better adaptability.Therefore,the expected effect is achieved.A support vector machine regression model based on genetic algorithm is established,which enriches the prediction method for short-term price of stock index futures.