Research on Stock Price Prediction Model based on GA Optimized SVM Parameters

This paper construct the predicted model based on support vector machine (SVM) for the Shanghai Composite Index, acquired the model parameters using genetic algorithm optimization was carried out, combined with k-fold cross method. Experiments based on the start date to February 2011 total 4948 trading day data, 10 fold cross circulation experiments of GA optimization; get the most accurate model parameter of SVM. At last, the regression model is used to predict, and the relative error of regression prediction is 0.11, and the accuracy of regression prediction is higher. In conclusion, this model can be used to predict the Shanghai Composite Index.