Deep convolutional fuzzy systems of stock value prediction based on AFS theory

The stock price market is affected by many factors, so it is very complex and difficult to predict the stock price. Using deep learning algorithms to analyze and predict the stock price data has a heavy computational burden and lacks interpretability for a large number of training parameters. In order to solve above problems, a new deep convolution fuzzy system called DCAFS is proposed, which combines the axiom fuzzy sets theory with the multi-level structure of deep convolution neural network to explore the multi-level expression ability of fuzzy rules and ensure prediction accuracy. This paper uses a chaotic time series to verify the feasibility of the system, and then DCAFS is used to predict the opening price of the stock market. The experimental results show that the DCAFS has a good prediction effect in complex time series analysis such as stock price.