Research And Application Of Stock Trend Prediction Based On Deep Neural Network

Stocks are a common financial product with high risks and high returns.In order to better analyze stock investment and obtain an effective stock selection plan,a model for predicting stock trends is proposed,and the improved model is applied to construct The stock trend prediction system enables investors to make relatively correct investment decisions based on the stock recommendation results and stock prediction results of the system,thereby effectively reducing investment risks and obtaining high and stable investment returns.Compared with the existing stock trend prediction model,the improved model combines stock price trends,news information and investor sentiment to make predictions.It not only uses the transaction data in the stock market,but also takes into account financial,political news and stock forum comments for the stock market.Impact.First of all,the sentiment classification of the stock forum speech is proposed,and a naive Bayesian-based stock forum speech model is proposed.Experiments have confirmed that the naive Bayes classifier has a better classification effect and can obtain investor sentiment more accurately.It laid the foundation for the establishment of the entire stock trend prediction model.Secondly,based on the original LSTM,this model is constructed by mixing BiLSTM and CLSTM.BiLSTM extracts stock transaction data and investor sentiment index related features.CLSTM integrates and processes the contextual characteristics of news,and finally outputs predictions through the fully connected layer.result.In the experimental model,the classification method is used to conduct experiments on stock trends,and the probability of being classified as stock rising and stock falling is obtained.The experiment used 300 shares of Shanghai and Shenzhen as data collection.The prediction effect is evaluated through the accuracy rate and the yield rate.The experimental results show that,compared with the single LSTM prediction model,the proposed method has a certain improvement in the accuracy rate of stock trend prediction,to a certain extent,it can accurately and effectively Forecast stock trends.At the same time,the stock trend prediction system of deep neural network is designed and implemented,and the trained prediction model is uploaded to the stock prediction module.Through the demand analysis of the stock trend prediction system and the design of various functional modules,the entire stock trend prediction system is completed through development,testing and other links.Investors can use this system for stock selection and investment.