Communications in Mathematical Finance

Stock Prediction via Linear Regression and BP Regression Network

  • Pdf Icon [ Download ]
  • Times downloaded: 9237
  • Abstract

    Stock prediction is an important and challenging problem. In this paper, two models, i. e., a multiple linear regression based model and a BP neural network based model were studied to predict the movements of the stock market. Through the analysis of the conceptual topic index of Shenzhen A-share “Gree Electric Appliances” (000651) and the prediction of stock price technical indicators on opening price of the next day, it was found that the results predicted by technical indicators on the opening price were more accurate than the conceptual topic index. At the same time, we compared the fitting degree, simulation and prediction ability of the multiple linear regression model and the BP neural network prediction model. The results show that the prediction model based on BP neural network works better than the multiple linear regression model. We trained and tested the index of “Gree Electric Appliances” in 220 trading days in this paper. Based on the experimental results, the conclusion was drawn that the BP neural network based model with the stock technical index as the characteristic variable can predict the movements of the stock market well.

    Keywords: BP neural network; linear regression; stock price prediction; R language