Stock forecasting based on finegrained evolution hypernetwork
1. College of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China; 2. Chongqing Key Laboratory of Computational Intelligence, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
Abstract：To predict the stock trend accurately, a novel stock forecasting method was proposed based on finegrained evolution hypernetwork. The securities annual report data in 2011 was processed to predict the trend of stock in 2012. To cope with the information loss by simple binarization of continuous data, the Chisquare splitting algorithm combined with equalwidth discretization method was used to discretize the dataset of stock. The finegrained evolution hypernetwork was adopted to predict the trend of each stock in the coming year. The experimental results show that the trend prediction accuracy of the proposed method for stock data is 86.73%, and the accuracy of stock rising forecasting is 75.00%. The important combination of features and the corresponding value range are obtained by the model impacting the future movement of stock, and a reliable and convenient method for stock selection is provided.
2017, Vol. 38 
