Amazing technological breakthrough possible @S-Logix

Office Address

  • #5, First Floor, 4th Street Dr. Subbarayan Nagar Kodambakkam, Chennai-600 024 Landmark : Samiyar Madam
  • +91- 81240 01111

Social List

A CNN-BiLSTM-AM method for stock price prediction - 2021

Research Area:  Machine Learning


In recent years, with the rapid development of the economy, more and more people begin to invest into the stock market. Accurately predicting the change of stock price can reduce the investment risk of stock investors and effectively improve the investment return. Due to the volatility characteristics of the stock market, stock price prediction is often a nonlinear time series prediction. Stock price is affected by many factors. It is difficult to predict through a simple model. Therefore, this paper proposes a CNN-BiLSTM-AM method to predict the stock closing price of the next day. This method is composed of convolutional neural networks (CNN), bi-directional long short-term Memory (BiLSTM), and attention mechanism (AM). CNN is used to extract the features of the input data. BiLSTM uses the extracted feature data to predict stock closing price of the next day. AM is used to capture the influence of feature states on the stock closing price at different times in the past to improve the prediction accuracy. In order to prove the effectiveness of this method, this method and other seven methods are used to predict the stock closing price of the next day for 1000 trading days of the Shanghai Composite Index. The results show that the performance of this method is the best, MAE and RMSE are the smallest (which are 21.952 and 31.694). R2 is the largest (its value is 0.9804). Compared with other methods, the CNN-BiLSTM-AM method is more suitable for the prediction of stock price and for providing a reliable way for investors to make stock investment decisions.

Author(s) Name:  Wenjie Lu, Jiazheng Li, Jingyang Wang & Lele Qin

Journal name:  Neural Computing and Applications

Conferrence name:  

Publisher name:  Springer

DOI:  10.1007/s00521-020-05532-z

Volume Information:  volume 33, pages 4741–4753 (2021)