Research on the Application of Machine Learning Algorithm in the Prediction of Stock Return
In view of the shortcomings of the accuracy and efficiency of the current stock return prediction method based on machine learning algorithm,a stock return prediction model is proposed.First of all,in view of the problem that the stock data contains more noise,the wavelet transform algorithm is used to denoise it.Then,based on the existing re-search results and the current situation of the stock market,the stock return prediction index system is constructed.Fi-nally,aiming at the defects of SVM,an improved Drosophila algorithm is used to optimize it,and a stock yield predic-tion model is constructed.The test results show that the error value of IFOA-SVM model is 0.18,the loss value is 0.23,the prediction accuracy is95%,and the AUC value is0.954.the method proposed in the study can accurately predict the stock return rate,thus improving investor returns and stimulating the development of the stock market.
NETL
NSTL
万方数据
