Robust Extension of FCMdd-based Linear Clustering for Relational Data using Alternative c -Means Criterion

摘要

Relational clustering is an extension of clustering for relational data. Fuzzy c-Medoids (FCMdd) based linear fuzzy clustering extracts intrinsic local linear substructures from relational data. However this linear clustering was affected by noise or outliers because of using Euclidean distance. Alternative Fuzzy c-Means (AFCM) is an extension of Fuzzy c-means, in which a modified distance measure based on the robust M-estimation concept can decrease the influence of noise or outliers more than the conventional Euclidean distance. In this paper, robust FCMdd-based linear clustering model is proposed in order to extract linear substructure from relational data including outliers, using a pseudo-M-estimation procedure with a weight function for the modified distance measure in AFCM.