Gaussian Collaborative Fuzzy C-Means Clustering

作     者：Gao, Yunlong Wang, Zhihao Li, Huidui Pan, Jinyan 

作者机构：Xiamen Univ Dept Automat Xiamen Peoples R China Jimei Univ Coll Informat Engineer Xiamen Peoples R China 

出 版 物：《INTERNATIONAL JOURNAL OF FUZZY SYSTEMS》 (Int. J. Fuzzy Syst.)

年 卷 期：2021年第23卷第7期

页      面：2218-2234页

核心收录：

主　　题：Fuzzy c-means (FCM) Gaussian collaborative Fuzzy clustering algorithms Gaussian mixture model (GMM) 

摘      要：For most FCM-based fuzzy clustering algorithms, several problems, such as noise, non-spherical clusters, and size-imbalanced clusters, are difficult to solve. Different fuzzy clustering algorithms are developed to deal with these problems from different perspectives. However, no comprehensive viewpoint to generalize these problems has been put forward. In this paper, we reveal the inherent deficiency of FCM and propose a new fuzzy clustering method called Gaussian Collaborative Fuzzy C-means (GCFCM) to solve these problems. In GCFCM, Gaussian mixture model (GMM) and collaborative technology are adopted to enhance the ability of recognizing the intrinsic structure of clusters. Experimental results confirm that GCFCM is effective in dealing with noise, non-spherical clusters, size-imbalanced clusters, and those also show excellent performance in dealing with real-world data sets.