Research On The Classified Prediction Of Stock Price Increase Based On Principal Component Analysis And Support Vector Machine

Stock price prediction has often been a subject of extensive research in the financial and academia field.The fall and rise of stocks directly reflects the return of investors.If the rise and fall of stocks can be predicted well,it can provide investors with a certain degree of auxiliary decision-making.Therefore,this paper takes the rise and fall of stocks as the research object.We always use support vector machines for classification problems and regression analysis.They have a good statistical theoretical foundation and strong model generalization ability.The principal component analysis method can extract the main original information and even achieve the optimal compression of samples.The PCA-SVM model doesn't change the distribution characteristics of samples,but also can improve the calculation speed.The development of the banking industry is more representative in the securities market because it is inseparable from the state of the country's economy.Therefore,this paper selects the stocks of the banking sector for prediction and analysis.The relevant 121-day stock data from August 20,2019 to February 21,2020 downloads on the Southwest Securities computer client,the tonghuashun computer client and the UQER database.The company's relevant financial indicators are mainly used to reflect the company's operating conditions better when performing cluster analysis.The trading indicator data can grasp stock trends better when forecasting gains.Use R software as the platform to analyze the data with the help of the self-editing program and the calling of related packages.First,the Net Inflow Open,Net Inflow Close,Net Rate XL and Main Rate are added to the selection of indicators so as to keep up with the main force and grasp the dynamic changes of stocks;Second,the weighted principal component clustering method is used to perform a cluster analysis on the stocks of the banking sector in order to understand and explore the correlation between various banks better.Third,classify the stock price increase according to the distribution of the stock price increase,and add the oscillation category as a buffer between rising and falling in the classification label.Fourth,cross-validation method and grid search method are used to find the optimal parameters and obtain the final classification model.According to the prediction results of the model,the investment success rates of China Merchants Bank and China Construction Bank are 85.37% and 87.8%.After the combined model prediction,the investment success rates of Zhengzhou Bank and Jiangyin Bank are 76.92% and 80.77%.The effect of PCA-SVM model is more significant compared with the Neural Network and Fischer discriminant method,and the average investment success rate exceeds 80%,which reflects the superior performance of the model in the classification prediction of the stock price increase of the banking sector.