Title: A neural network-based ensemble forecasting method for financial market prediction

Authors: Wei Xu; Meiyun Zuo; Mingtao Zhang; Rong He

Addresses: School of Information, Renmin University of China, Beijing, 100872, China. ' School of Information, Renmin University of China, Beijing, 100872, China. ' School of Information, Renmin University of China, Beijing, 100872, China. ' School of Information, Renmin University of China, Beijing, 100872, China

Abstract: This paper presents a neural network-based ensemble forecasting method for financial market prediction. In the proposed approach, each of the predictors is firstly constructed by training on a set of samples produced by bootstrapping using neural networks. Then, the constraint conditions are offered to select competitive predictors in order to improve the forecast performance. Finally, with the proper predictors, several ensemble strategies are suggested to combine the results of single predictors. During the process, the network structures of neural networks are discussed, and the optimal network structure and appropriate parameters of the proposed model are determined by grid search method. To validate the proposed method, stock price data is used for evaluation. The results show that the neural network-based ensemble method outperforms traditional bagging forecasting method and single neural network predictor. These findings imply that the proposed method is a promising approach for financial market prediction.

Keywords: financial markets; stock prices; neural networks; ensemble method; market prediction; forecasting; stock markets.

DOI: 10.1504/IJAMECHS.2011.043374

International Journal of Advanced Mechatronic Systems, 2011 Vol.3 No.4, pp.259 - 267

Published online: 18 Mar 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article