Title: Type-1 fuzzy time series function method based on binary particle swarm optimisation

Authors: Cagdas Hakan Aladag; Ufuk Yolcu; Erol Egrioglu; I. Burhan Turksen

Addresses: Department of Statistics, Faculty of Science, Hacettepe University, 06800, Ankara, Turkey ' Department of Statistics, Faculty of Science, Ankara University, Ankara, Turkey ' Department of Statistics, Faculty of Science, Ondokuz Mayis University, Samsun, Turkey ' Industrial Engineering Department, TOBB Economy and Technology University, Ankara, Turkey

Abstract: For time series forecasting four kinds of fuzzy-based approaches can be used. These are fuzzy regression techniques, fuzzy time series methods, fuzzy inference systems, and fuzzy function approaches. There are some major problems in using fuzzy regression techniques and fuzzy inference systems for time series forecasting. Therefore, it would be wise to use a forecasting approach which combines fuzzy time series and fuzzy function approaches. In this study, a fuzzy time series forecasting method based on fuzzy function approach is proposed by adopting fuzzy function approach to time series forecasting. And, the proposed approach is called type-1 fuzzy time series function approach. Also, in the proposed approach, the lagged variables of the system are determined by using binary particle swarm optimisation. In order to evaluate the performance of the proposed method, it has been applied to well-known time series of and Istanbul stock exchange dataset.

Keywords: time series forecasting; fuzzy time series; fuzzy functions; binary PSO; particle swarm optimisation; Australia; beer consumption; Turkey; stock markets; fuzzy logic.

DOI: 10.1504/IJDATS.2016.075970

International Journal of Data Analysis Techniques and Strategies, 2016 Vol.8 No.1, pp.2 - 13

Published online: 20 Apr 2016 *

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