Title: A hybrid approach to forecast stock market index

Authors: Gurbinder Kaur; Joydip Dhar; Rangan K. Guha

Addresses: Sant Longowal Institute of Engineering and Technology, Longowal-148106, Punjab, India ' ABV-Indian Institute of Information Technology and Management, Gwalior-474015, MP, India ' Sant Longowal Institute of Engineering and Technology, Longowal-148106, Punjab, India

Abstract: The forecasting of stock market problem from the available data is quite often of uncertain nature, hence the stock market prediction is a very challenging and difficult task. In this paper, we have investigated the predictability of stock market of Bombay Stock Exchange (BSE30), Hang Sang China Stock Index (HS), Japan Stock Index (NIKKEI) and Taiwan Weighted Index (TWI) with adaptive network-based fuzzy inference system (ANFIS) combined with subtractive clustering technique. In this process, we compared stock markets with variable numbers of data clusters. Optimised subtractive clustering is used to cluster the data and create fuzzy membership functions. Finally, a hybrid learning algorithm has been used to combine least square method and back propagation gradient-decent method for training the fuzzy inference system. This paper represents a state of the art for ANFIS application to forecast stock market index.

Keywords: neural networks; fuzzy logic; stock market forecasting; stock markets; subtractive clustering; stock market predictability; adaptive neuro-fuzzy inference system; adaptive network-based fuzzy inference system; ANFIS.

DOI: 10.1504/IJAISC.2015.070638

International Journal of Artificial Intelligence and Soft Computing, 2015 Vol.5 No.2, pp.165 - 176

Accepted: 08 Mar 2015
Published online: 15 Jul 2015 *

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