Research On Neural Network News And Volume-price Modeling For Stock Price Prediction

Stock price prediction has important significance in the business and financial fields.In recent years,the use of artificial intelligence technology for data analysis and prediction in the financial sector has become a hot topic.In the direction of stock price forecasting,the traditional method is to analyze the historical data of stock prices.With the advent of the era of big data,more and more people are considering using news and information as a basis for stock price prediction.This paper proposes a hierarchical modeling method based on hierarchical neural networks.It predicts stock price trend by modeling stock-related news on the Internet.The main research content includes:(1)Word2vec is used to transform news vocabulary into word vector,hierarchical neural network based on recurrent neural network and convolutional neural network is used to semantically model the news text,and news headline information and attention mechanism are introduced to give different news text sequences.The degree of concern ensures that the model can capture information more relevant to changes in the stock price trend from the news;(2)Using feature engineering methods,features are extracted from the stock history volume data,normalized and then modeled using multi-layer neural networks to model volume and price,and then combined with modeling news results to jointly forecast stock prices.(3)A stock simulation trading system was designed to build three test sets: bull market,bear market,and plain market based on the overall performance of the stock market.The confidence of the stock was predicted based on the prediction model,and the stock trading was performed through the simulation board to calculate the investment obtained.Benefits to evaluate the performance of the forecast model.The model we eventually get can use stock-related news and historical stock price data to give a confidence level to stocks every day.Based on the confidence level,it can guide the trading behaviors of securities investors and help them to increase the returns from investing in the stock market.