Stock Price Prediction And Improvement Based On Stock Technical Indicators And Company News

The stock market is one of the important components of a country’s economic market.Compared with some capitalist countries such as Europe and the United States,my country’s stock market has a shorter development time and an immature system.With the continuous development of my country’s economy since the reform and opening up,the continuous improvement of people’s living standards and the continuous accumulation of wealth,the stock market is not only a wealth growth channel for investors,but also more and more ordinary people participate in the stock market.However,in the complicated stock market,how to choose investable stocks has become a problem to be solved.In order to avoid investment losses,the trend prediction of stock prices has become an important hot issue and difficult problem.Therefore,the design and implementation of the stock market forecasting system not only has profound theoretical significance,but also has important practical value.Therefore,this paper uses the xgboost model to predict the price of stocks based on the technical indicators of stocks and the news of listed companies,and finally uses the grid search method to improve the model to improve the accuracy of the forecast.This paper firstly summarizes the research status of stock price forecasting at home and abroad,and introduces the model theory used in this paper in detail,including xgboost model theory,principal component analysis and stock market overview.Next,this paper selects five stocks of iFLYTEK,Ping An,PetroChina,CRRC and Sany Heavy Industry for empirical analysis.First,the historical transaction data and stock technical indicators of the above five stocks are obtained.Through principal component analysis reduce the dimension of the sample data,and use the reduced data as the input feature of the xgboost model to predict the stock price.The prediction results show that the use of the xgboost model based on stock technical indicators can roughly predict the trend of the stock price,but the difference between the actual value and the predicted value.There is a big difference between them.Then,in order to improve the accuracy of prediction,this paper obtains the daily news headlines of listed companies,scores their sentiments through text analysis,gets the daily news sentiment scores,and adds them to the input features of the model to predict the stock price.,and found that the accuracy of stock price forecasts has been improved to a certain extent.Finally,in order to improve the prediction performance of the model,this paper adjusts and optimizes the more important 8 parameters in the xgboost model using the grid search method,finds the best parameters,and uses the parameter-optimized model to predict the stock price.The model adjusted by the search method can greatly improve the accuracy of the model’s stock price prediction.After demonstration and analysis,the possible innovations of this paper are as follows:(1)Perform sentiment analysis on the daily news of listed companies,calculate their sentiment scores,and then add them as features to the training set to predict stock prices,so as to predict the stock price.To explore the impact of news sentiment on the accuracy of stock price forecasts.(2)The xgboost financial forecasting model optimized by grid search method is proposed.In this paper,8 important parameters in the xgboost model are selected,and the grid search method is used to adjust and optimize them.