Research Of Stock Price Prediction Based On SVR With Parameters Optimized By Improved GA

Stock price is a non-static and changeful system. Many factors can influence the stock price and the change of price fluctuations is complicated. Therefore, to effectively predict the stock price, avoid this high-risk of stock risks and reap profits become the research direction of the international and domestic academics today.Aiming to the dynamics and nonlinearities of stock price, a stock price prediction model that based on the improved genetic algorithm (GA) optimizing the parameters of support vector regression (SVR) was proposed. First, the wavelet was used to de-noise the samples of stock price. The wavelet can eliminate the noise in the stock price time series, and fully retain effective features of the original signals, greatly improving the accuracy of forecasts. Then the SVR model whose parameters were optimized by the improved GA was utilized to predict and assess the data de-noised by wavelet. The result demonstrated that the improved wavelet-GA-SVR model has good prediction effect compared to the standard wavelet-GA-SVR, the single SVR and other prediction models, and it is significant to the study of the prediction of stock price.