Research On Stock Price Prediction Based On Deep Convolutional Fuzzy System Model

Economic forecasting is an eternal subject that economists explore,but it is a challenging task.Among them,the stock market,as a "barometer" reflecting a country's macro economy,plays a crucial supporting role in the national economy.The prediction of stock price has always been the focus of many scholars and investors,so it is of great significance to study the predictability of stock price.This paper compares and analyzes two different viewpoints of predictability and unpredictability of stock price by combing domestic and foreign literatures,and studies the influence of investors' behavior on stock price and predicts the changes of stock price from the perspective of predictability of stock price.This paper analyzes the performance characteristics of stock prices based on the classical theories related to behavioral finance and combining with the technical analysis theory and chaos fractal theory.Combining with the technical analysis method,this paper analyzes the technical analysis indexes related to the changes of stock prices,and analyzes the relationship between each technical analysis index and investor behavior one by one.Secondly,based on the technical analysis index of price and quantity,this paper USES the Laplace scoring algorithm to construct the index of investor behavior.Finally,with the concept of quantitative analysis,the deep convolutional fuzzy system model(DCFS)is used to predict the stock price with the constructed investor behavior index,and the prediction performance of the investor behavior index is tested according to the prediction results.In this paper,the technical analysis indexes of the daily opening price,closing price,maximum price,minimum price and equivalent volume of were selected as sample data for empirical research.Finally,the following conclusions are drawn :(1)the technical analysis index of price and quantity is effective and can better reflect the change of stock price.(2)investor behavior indicators can be used to better predict the trend of stock price changes.(3)the efficient market hypothesis is not completely valid,and there is an indicator variable in the stock price that can be used to predict the stock price.Shares due to objective factors lead to instability model parameters and the influence of heterogeneity,this article to add the appropriate price data noise is used to replace the objective factors,and then reproduce the real market cannot be observed the influence of factors on the stock price changes,and the substitution of data driven analysis model for technical analysis method of graphical analysis model of stock price forecasting analysis,stock price change rule by careful analysis of the quantitative concept investigation.