The Application Of Particle Swarm Optimization In Fuzzy Time Series Prediction Problem

In the current treatment of the study of mathematics, forecasting problem based on the concept of fuzzy time series has become an important research topic. To improve the forecasting accuracy of fuzzy time series, many scholars presented all kinds of fuzzy forecasting method, the research focus has two aspects:the lengths of intervals and the content of the rules.In order to improve the prediction accuracy, I-Hong Kuo, et al (1995) proposed using particle swarm optimization algorithm to deal with interval partitioning of fuzzy time series prediction problem. The main idea is:every particle defines a n-1 d vector, the vector will be divided domain into n interval, then a group of particles move in the scope of the universe of discourse, find out optimal division. I-Hong Kuo presented a new fuzzy forecasting method (named MV). The prediction results of forecasting enrollments of students of the University of Alabama show that the model is better than the previous algorithm.The fuzzy set of previous given is used in I-Hong Kuo’s paper, this method of defining fuzzy set is cannot reflect the real distribution of data structure, so Professor Chen Gang proposed a objective definition of fuzzy set--istance definition, it avoid the disadvantages by subjective defining of previous fuzzy sets. Fuzzy set of distance definitior is used in this paper, and the fuzzy forecasting method is to make a little change based on MV method. The last results show that forecasting accuracy of this algorithm is better than algorithm of I-Hong Kuo.