Prediction Of Stock Price Using Gaussian Process Model

The stock market in our country has experienced thirty years’ development and revolution,it has made great contribution to the economic construction and social development and has become one of the most important industries in our country economy.The stock market working efficiency will directly decide the financing capacity of the listed enterprises and the market active level of the fund surplus ones,so it’s development and changes.is closely related to the interests of enterprises and investors.And stock price can best reflects fluctuations and changes of stock market,so the prediction of stock price has been the issues which the investors are always concerned about.This paper focuses on the research of the closing price data of Sinopec stock.First,the traditional ARIMA model is used for the prediction of the stock price time series data,the model has a certain degree of limitation.Then introduce the Gaussian Process Model to predict stock price,by using the cross validation method,the original data set is divided into training set and test set,the optimum kernel function and the optimal hyperparameter are selected by training the training set,the optimal fitting model is obtained,and through the test set to evaluate the model and forecast research,a considerable prediction effect is obtained.The main highlight of this paper is introducing the Gaussian process model to the stock market,the Gaussian process model is one of machine learning algorithms,the article demonstrates the feasibility of using the Gaussian process model to predict the stock price trend,theoretically also further promote the application of machine learning method in the stock market.