Research And Application Of Stock Index Pre-model

There are many participants in the stock market,the influencing factors are complex,the transaction data is time-sensitive,and the noise interference is high.On the basis of complex financial data,how to effectively judge the state of the stock market and the future trend of change have important practical significance for market participants and managers.In this paper,based on the support vector machine model,the differential evolution algorithm is used to optimize its parameters.Based on this,two models are built to predict the complex stock index.One is based on the ARMA-SVM model.The model deeply integrates the support vector machine with the time series model,and references the variance sequence obtained by the GARCH model to the model to improve the prediction accuracy,which not only retains the advantages of the support vector machine,but also It can better reflect the timing characteristics of the sequence.The other is a stock index forecasting model based on mutual information and RBM feature extraction.The model makes full use of all levels of information to integrate technical indicators and surrounding markets into the model.The mutual trust information is used to extract features from the RBM stacking deep trust network.Evolutionary algorithm optimized support vector machine for prediction.The example verification of the Shanghai and Shenzhen 300 Index and the HengSheng Index shows that the two models constructed in this paper have higher prediction accuracy and can provide some help for investor market analysis and decision making.