Research And Implementation Of Stock Forecasting Model Based On Neural Network

The stock market is a barometer of the national economy,which can reflect the stability and health of the economy to a certain extent,so it has an important position in the national economic system.How to accurately predict the stock price and trend is an important problem for the industry and academia.Artificial intelligence technology,represented by deep learning,is currently being widely applied in various fields,and has broad application prospects in the financial field.Deep learning has the ability to learn data rules from massive data.This paper draws on the idea of using indicators to predict prices in technical analysis,and designs a stock prediction model based on deep neural network.The model consists of two parts: feature extraction part and stock prediction part.The feature extraction model is based on deep self-encoders,which is used to extract features from stock index data to represent the current state of stocks.The prediction model can be constructed according to the long time memory neural network,and predict the current price according to the historical state.Finally,the dropout and L2 regularization methods are used to optimize the model to reduce the risk of over fitting.In this paper,we collected data from the wine making section and built data sets.We designed a number of experiments to verify the effectiveness of the model.Firstly,the self-encoders of different depths are constructed to predict the accuracy and restructure the error to compare with the coding effect.Secondly,the extraction results of this model and other feature extraction methods are compared.Principal component analysis(PCA)and factor analysis(FA)were used as comparison objects to analyze the differences between different methods.Finally,the prediction effect and running time of the model and other prediction models are compared.The experimental results show that the model designed in this paper has better extraction effect,higher prediction accuracy,moderate time complexity and practicabilityFinally,this paper uses Python to design a set of stock forecasting system based on B/S architecture.This system has realized the function of the prediction result display,the real-time stock data acquisition and so on,which provides the convenience for the users.