Active Learning of constraints using incremental approach in semi-supervised clustering

Semi-supervised clustering aims to improve clustering performance by considering user-provided side information in the form of pairwise constraints. We study the active learning problem of selecting must-link and cannot-link pairwise constraints for semi-supervised clustering. We consider active learning in an iterative framework; each iteration queries are selected based on the current clustering outcome and constraints available. We use the neighborhood framework where the pairwise points having the must-link belong to the same neigborhood and cannot-link pairwise points belong to the different neighborhood. If two points belong to the same neighborhood then they belong to the same cluster and viceversa. We will use the Glass Identification Data Set from the UCI machine learning repositories and investigate the improvement in clustering time using the Incremental Clustering. Keywords— Active learning, Semi-supervised clustering, Incremental Approach, Pairwise constraints.