Stock price prediction based on stock price synchronicity and deep learning

Abstract

Deep learning technology has been widely used in the financial industry, primarily for improving financial time series prediction based on stock prices. To solve the problem of low fitting and poor accuracy in traditional stock price prediction models, this paper proposes a stock price prediction model based on stock price synchronicity and deep learning methods, which applied the stock price synchronicity theory in stock price trend analysis. This paper first uses the affinity propagation algorithm to build stock clusters, and then, based on convolution neural network (CNN), and feature weight to construct the stock price synchronicity factor. At last, the long short-term memory (LSTM) network with multifactor is built for stock price trend analysis. According to the theory of stock price synchronicity, the affinity propagation algorithm can find the potential related stocks of the target stock. The spatial data analysis ability of the CNN model provides a guarantee for the application in stock price synchronicity factor analysis. The LSTM model can better analyze the information contained in the stock price time series and predict the future price. The experimental results show that, compared with the traditional multilayer neural network model, the LSTM model has better accuracy in the trend prediction of the stock price. Simultaneously, the application of stock price synchronicity effectively improves the performance of the multifactor LSTM network.