Fuzzy clustering to classify several regression models with fractional Brownian motion errors
Mohammad Reza Mahmoudi, Mohammad Hossein Heydari, Kim-Hung Pho
Abstract
Clustering regression models fitted on the dataset is one of the most ubiquitous issues in different fields of sciences. In this research, fuzzy clustering method is used to cluster regression models with fractional Brownian motion errors that can be fitted on a dataset. Thereafter the performance of proposed approach is studied in simulated and real situations. The results verify that the introduced technique has excellent power to cluster the models. It indicates that our proposed method obtain many advantages. The performance of proposed technique is allowable. In addition, the algorithm is not so complicated. Furthermore, this method can be employed to compare both linear and nonlinear models.