Research On The Prediction Model Of Stock Price Based On SVM

As China enters the 21 st century,Chinese economy is developing rapidly,and the financial market is becoming more and more standardized.The stock market is an important part of the financial industry.Using models to accurately predict the trend of stock prices has important practical significance for investors to formulate investment strategies and avoid market risks.Therefore,more and more scholars are devoted to related research on the stock market.As stock prices are affected by many aspects,the stock market is becoming more and more complex.Simple time series models cannot solve the non-stationarity and nonlinearity of stock data.Therefore,the shortcomings of traditional linear models are gradually exposed.With the advent of the era of big data,machine learning technology has gradually been applied to stock research,allowing us to have more methods to predict stock prices.The typical methods are BP neural network and support vector machine model.Through the analysis of the stock market,we select the Shanghai stock index and five industry representative stocks as the original data of this paper,and select eight main influencing factors as the input variables.First,we use principal component analysis to preprocess the data and extract the effective information to the maximum extent.In order to judge the prediction results,we select three evaluation indexes,namely RMSE,Mae and the accuracy of rise and fall.Then we use support vector machine classification model and support vector machine regression model to predict the stock trend and stock closing price respectively.The results show that the prediction results of the support vector machine model after data preprocessing have better prediction effect than the original data.In order to he accompare tcuracy of the model,the classical BP neural network and the traditional ARIMA model are established.The results show that the support vector machine has a good prediction effect.