TECHCOMB

Volume 2 : Issue 2

A Fuzzy Time Series Approach Using De/Best/1 Mutation Strategy of Differential Evolution Algorithm

Authors : Eren Bas, VedideRezanUslu, UfukYolcu, ErolEgrioglu

image

Abstract:

Fuzzy time series approaches have been frequently used in recent years because of not need the assumptions valid for traditional time series. Fuzzy time series compose of three stages. These stages are fuzzification, determination of fuzzy relations and defuzzification stages, respectively. In the literature, there are many studies on each stage. Artificial intelligence optimization algorithms have been used frequently in almost all areas in recent years and also they have been used in different stages of fuzzy time series approaches. Genetic algorithm, particle swarm optimization and differential evolution algorithm are the most popular algorithms among artificial intelligence optimization algorithms. In this study, we proposed an approach to determine the sub-intervals by using differential evolution algorithm and also we used the DE/Best/1 Mutation Strategy in the stage of mutation which is the important stage of differential evolution algorithm to obtain better forecasting results. By using DE/best/1 mutation strategy it is aim that the system is rescued from randomness and the best chromosome is constantly kept in the system by updating its each iteration. Then, this proposed approach was applied to two real data set and the obtained results were compared with the other results suggested in the literature.