期刊名称：International Journal of Computer Science Issues
印刷版ISSN：1694-0784
电子版ISSN：1694-0814
出版年度：2012
卷号：9
期号：3
出版社：IJCSI Press
摘要：Many researchers have used fuzzy logic system to predict the time series data. In fuzzy system, the crisp data are converting into fuzzy data based on membership function. The futuristic data is predicted using previous data and fuzzy relation. But, in fuzzy system, there are many existing and derived membership functions which are used to fuzzify data. In this paper, an effort has been made to predict the time series data based on different fuzzy membership functions like Gaussian, Triangular, S-function, trapezoidal, Gbell, Dsigmoidal, Psigmoidal and Pi-shaped. A comparison has been made among the predicted data using different membership functions. One membership function has been selected based on minimum error in prediction of data. This process has been repeated on fifteen time series data sets. Finally, one membership function has been selected which has given minimum error in maximum cases.
关键词：Many researchers have used fuzzy logic system to predict the time series data. In fuzzy system; the crisp data are converting into fuzzy data based on membership function. The futuristic data is predicted using previous data and fuzzy relation. But; in fuzzy system; there are many existing and deriv
