Design And Implement Of Stock Price Assistant Prediction System Based On Company Report Text Mining

With the gradual improvement of people's living standards,a series of changes have taken place in the way people invest and manage money.More people are gradually paying attention and participating in the stock investment market.Stocks have high returns and are accompanied by higher risks.The changes in stock prices are affected by many factors.Therefore,research on stock price prediction is of great value.The prediction of stock prices has become a problem that many scholars and stockholders continue to explore since the emergence of the stock market.In recent years,artificial intelligence technology has been widely used in different fields,and has achieved remarkable results in computer vision,speech recognition and natural language processing.The organic integration of machine learning and stock forecasting is a major trend in the field of financial stock forecasting.The LSTM model has a better performance in analyzing and forecasting based on historical stock prices,but its training data set is pure stock historical trading data.Such digital information content is thin,and there is much room for improvement in the performance of the prediction model.The report issued by a listed company,as a summary of the company's operation at a certain stage,can greatly reflect the company's capital,profit status and current status of the company's viability.It is also the focus of the market and has a great impact on the subsequent trend of the stock.Therefore,this study uses natural language processing methods based on machine learning,uses word vector technology to structure unstructured text information in listed company reports,and uses machine learning-related algorithms to find the internal relationship between company reports and company stock prices.Get the possibility that the text of the company report will raise the stock price.Use this probability data as a dimension in a set of input data for time series-based stock predictions.At the same time,combining the advantages of long-term and short-term memory neural networks in historical stock price analysis and the advantages of machine learning and deep learning to process text,a stock price prediction model that can use time series analysis of effective information reported by listed companies is established.By studying and mining the historical report texts of listed companies,predicting the possibility of the company's stock price rising after the real-time report is issued,establishing a text prediction model that can make full use of the market's known information,and using its prediction results as one based on historical stock price prediction Impact factor,establish the LSTM stock prediction model with text mining as an auxiliary role,and realize the stock analysis and prediction system.Due to the complexity and uncertainty of the financial market,this forecasting method considers that changes in stock prices are the result of multiple events and multiple factors.Connect the company's operating status with the company's stock,and improve the performance of stock prediction through effective text information.