Prediction And Improvement Of Stock Market Model Based On Grey System Theory

With the development of society and the progress of science,people's material standard of living is improving day by day,and the stock investment has gradually entered into thousands of households.Because the stock investment has the characteristic of high income,it makes people rush for it,but with the high income of the stock,the high risk of the stock can not be ignored.Therefore,how to use the existing stock market information to predict the stock market changes,as far as possible to avoid risks and improve returns,has become an important issue.Based on the grey system theory and the support vector machine theory,this paper studies the periodical data of Shanghai Composite index.As an intrinsic energy system,the stock market includes a large number of historical data,whose characteristics of highly nonlinear make the linear regression method commonly used ineffective,so we consider using boundary value containing the modified GM(1,1)model changes on the stock market for prediction.In the forecast,we are able to give predicted level and determine whether the prediction is reasonable according to the grade the accuracy evaluation of the existing standard.In view of the error of the grey model is sometimes too large,we' d like to use the machine error correction model to improve the prediction accuracy via support vector regression after obtaining the residual sequence.In the support vector regression machine,we can improve the accuracy of the model by adjusting the penalty factor C of the SVR and the kernel function parameter gamma.By the improved GM(1,1)forecasting results of the new model,we also carry on the evaluation and comparison with the original model,found that the evaluation indexes are improved,so as to achieve the purpose to improve the original model.