Medium-and Long-Term Stock Price Trend Prediction Based On Knowledge Graph

As an important embodiment of measuring the economic health of a society and a country,and even the development of the industry,the state of the stock market is regarded as the most "fashionable" and the most accepted investment mode by the public because of its low threshold and simple operation.It used to be an important conversation for everyone after dinner.In the past when big data was not so popular and transparent,people liked to buy stocks through so-called "gossip" and speculation,and the results were mostly determined by "luck".Now,with almost everyone having economic common sense and big data all over the life track,relying on stock investment has become a technical means,instead of "luck" speculation,but most investors do not have the professional level of professional investors and cannot reasonably value the stock price,which may directly or indirectly lead to the instability of the stock market and even affect the healthy development of the national economy.In the existing research on stock price prediction,most of them use stock indicators to predict the real-time price of stocks or use stock public opinion data to predict stocks.These research directions are more short-term,and the factors considered are relatively single.From a practical point of view,there are many factors that affect the change of stock price.At the same time,different factors will have different response time and time effect on the impact of stock price.Therefore,stock should be used as a long-term investment tool rather than a "speculative" tool.In order to realize the medium and long-term prediction of stock price and let stock investors make more accurate judgments on stock valuation,The purpose of this paper is to construct the stock domain knowledge graph by combining the related technologies of knowledge graph,integrate the rich semantic information in the stock knowledge into the stock news events,and take the continuous stock news events based on knowledge graph as one of the variables affecting the change of stock price,At the same time,various financial indicators and patented technology are used as auxiliary variables to predict the stock price.Finally,the range of medium and long-term price rise and fall is predicted by constructing a multi input neural network model.Through the comparative test,the results show that the medium and long-term price trend prediction algorithm based on knowledge graph proposed in this paper improves the prediction accuracy and recall rate to a certain extent compared with the compared algorithm.