An improved Fuzzy Time Series forecasting model based on Particle Swarm intervalization

Abstract

The objective of this paper is to show the strength of a modified version of Particle Swarm Optimization (PSO) in definition of suitable partitions of fuzzy time series forecasting and increasing its accuracy. Although a lot of contributions have been made to increase the quality of forecasts using Fuzzy Time Series during recent years, there are only a few papers considering tuning the length of intervals in forecasting. In this paper, we propose a new method to tune the length of forecasting intervals and show the superiority of our procedure to those previously proposed using the well-known data of University of Alabama. The main contribution of this paper is to use a modified and effective PSO algorithm in which velocities are updated using a modified version of traditional PSO in order to have some diversification in solutions generated. In addition, monotonically decreasing functions for PSO parameters are used to improve the accuracy of forecast. The results show that our model outperforms other methods in the literature.

A combined robust fuzzy time series method for prediction of time series

“…Optimization of ratio has been put forward by Yolcu et al [97]. Recently, some heuristic optimization algorithms have been used to get better results by Kuo et al [65,66], Davari, et al [43], Park et al [77] and Hsu et al [56], Chen and Chung [29], Lee et al [69,70]. Chen [20] and Chen and Chen [21] preferred to use entropy-based partitioning approach to get intervals in his high-order forecasting model.…”

An ARMA Type Fuzzy Time Series Forecasting Method Based on Particle Swarm Optimization

“…In a few fuzzy time series studies, particle swarm optimization method has been exploited in fuzzification phase. While the particle swarm optimization method was employed by Davari et al [12] for fuzzification in the firstorder fuzzy time series forecasting model, Kuo et al [14] utilized the method in high-order models. In the fuzzification phase, Kuo et al [15] utilized the particle swarm optimization method in both the first-and the high-order models.…”

“…In the literature, various methods have been proposed to fuzzify observations. While fixed interval lengths are used in Song and Chissom [1][2][3], Chen [4], Huarng [5], Chen [6], Tsaur et al [7], Singh [8], and Egrioglu et al [9,10], dynamic length of interval lengths is employed in Huarng and Yu [11], Davari et al [12], Yolcu et al [13], Kuo et al [14,15], Park et al [16], Hsu et al [17], and Huang et al [18] in order to partition the universe of discourse. Also, Cheng et al [19], Li et al [20], Egrioglu et al [21], Chen and Tanuwijaya [22], and Bang and Lee [23] used some methods based on clustering algorithms.…”

High order fuzzy time series forecasting method based on an intersection operation

“…Yolcu et al [24] used a single-variable constrained optimization to determine the ratio for the length of intervals. Moreover, to get the intervals as dynamic, while Davari et al [25], Kuo et al [26,27], Park et al [28], Hsu et al [29], Fu et al [30], and Huang et al [31] took advantage of particle swarm optimization; Chen and Chung [32] and Lee et al [33,34] benefited from genetic algorithm.. Even though some of these approaches determine sub-intervals by avoiding subjective judgments, the membership values are not still specified objectively .The approaches using fuzzy clustering techniques in fuzzification step have been proposed to get over this problem.…”