Multi-View Fuzzy Clustering Based on Adaptive Local Neighbor Information

Multi-view fuzzy clustering methods provide numerous solutions for clustering multi-view data. However, most existing fuzzy clustering methods primarily determine the membership of samples to different clusters based on the similarity between each sample and the cluster centers under various views, neglecting the similarity between samples and data within different clusters, and failing to effectively distinguish outliers. To address these issues, we propose an adaptive local neighbor information-based multi-view fuzzy clustering algorithm (ALN-MVFC). This algorithm considers the adaptive local neighbor information of each sample in different clusters, where the membership of a sample is determined by both its similarity to the cluster center and its similarity to several nearest neighbors from different clusters. Additionally, we introduce a novel outlier detection method and a way to identify the varying importance of different views. Experiments conducted on four widely-used datasets demonstrate the superiority of our algorithm.