Stock Price Prediction Based On Convolutional Neural Network

In recent years,the world economy has developed rapidly,at the same time,the uncertainty of the changing trend of financial industry has also increased.To research the trend of financial industry by studying the laws of financial activities is a very worthy subject and an important basis for making financial plans and decisions.In the study of financial time series,stock price prediction has always been a difficult point,but also a hot spot.The traditional method of stock price forecasting takes into account the macroeconomic situation and the development of enterprises.However,with the development of society and industry,traditional methods have become more complex and time-consuming.With the development of artificial intelligence,the method of applying machine learning to stock research is timely.The application of neural network in the financial field is a big climax of stock price prediction.convolution neural network(Convolutional Neural Networks)is a special multi-layer neural network,which is mainly composed of input layer,convolution layer,pool layer,full connection layer and output layer.it can be regarded as a feature extractor with excellent performance and high degree of automation program.At present,CNN are widely used in speech recognition,image classification etc.Accurate and timely financial prediction is the scientific basis for studying and judging financial forms,formulating financial strategies and making financial decisions.Without forecasts,financial decisions lack the necessary conditions.Therefore,financial forecasting plays an important role in both national economic development and financial enterprise management.This paper first summarizes the research methods of machine learning in stock price prediction at home and abroad,puts forward the significance of using convolution neural network to predict stock price,focuses on thestructure and principle of convolution neural network and the method of constructing convolution neural network model,and then combines stock prediction with artificial intelligence.Based on the MXNet neural network platform,this paper constructs a convolutional neural network model,combines the stock data of Shanghai Stock Exchange,uses its powerful supervisory learning ability and the unique ability of convolutional neural network to extract features,learns the features in financial data,and obtains that it can be pre-determined The model of stock price measurement,compared with different evaluation indexes,has finally achieved more satisfactory results,which provides a strong basis for the formulation of subsequent investment strategies.Considering that the stock price forecast is based on the "four-price one quantity" in the past,and then according to the characteristics of financial data,combined with the average trend(ADX)index,the model is improved,and the model is optimized from many angles.degree,exert greater learning ability.