Adaptive Fuzzy C-Means with Graph Embedding

• DOI: 10.48550/arxiv.2405.13427
• Publication Date: 2024-05-22
• Authors: Qiang Chen; Weizhong Yu; Feiping Nie; Xuelong Li
• Keywords: Graph Embedding

Fuzzy clustering algorithms can be roughly categorized into two main groups: C-Means (FCM) based methods and mixture model methods. However, for almost all existing FCM methods, how to automatically selecting proper membership degree hyper-parameter values remains a challenging unsolved problem. Mixture while circumventing the difficulty of manually adjusting hyper-parameters inherent in often have preference specific distributions, such as Gaussian distribution. In this paper, we propose novel that is capable learning an appropriate value handling data with non-Gaussian clusters. Moreover, by removing graph embedding regularization, proposed degenerate simplified generalized model. Therefore, also seen embedding. Extensive experiments are conducted on both synthetic real-world datasets demonstrate effectiveness