Stock Price Prediction Based On PCA-SVM-KNN Model

The stock market plays an important role in the economic development of a country.The prediction of stock price changes is very important for government departments,investment institutions and investors.However,the price trend of the stock market may be affected by political factors,macroeconomic factors,legal factors and other factors,resulting in great uncertainty in the study of stock price volatility and stock forecasting.How to select useful information from many factors to predict the stock has been widely studied by many experts and scholars.In the first part of this paper,the application of support vector machine in stock forecasting is studied,and a PCA-SVM-KNN model is proposed to predict the rise and fall of stock price.The principal component analysis is used to reduce the dimension of the selected stock basic data,so as to eliminate the multicollinearity among variables and extract the main components that affect the stock price,so as to reduce the running time of the model and improve the accuracy of the model prediction.The mixed model of support vector machine and k-nearest neighbor is improved,in which different time sliding windows are selected to classify and predict the stock trend.In the empirical study,Shanghai index and Shenzhen index are selected,and 12 stocks are selected from different industries.The results show that the proposed PCA-SVM-KNN model has great generalization ability and can accurately predict the trend of stock price.The second part of this paper is to predict the stock return and find the optimal portfolio.Firstly,considering the advantages of wavelet de-noising in processing nonstationary data in time prediction,the wavelet analysis method is combined with time series.The hybrid model is used to predict the stock return accurately.The fitting effect and prediction accuracy of the model are used to measure the performance of the model.At the same time of model optimization,wavelet transform and time series model also have some advantages in short-term forecasting.On the other hand,it also enriches the stock forecasting system.Finally,the optimal portfolio of the model is given when the condition of maximum return is satisfied.Through the research of this paper,I hope the hybrid forecasting model can bring some reference to the stock forecasting research and investors.