Constrained Clustering Based on Semantic Information
Takahiro Nishigaki, Takashi Onoda · 2013
Many clustering methods have been proposed due to the difference of the rule to generate a cluster. But the clustering is unsupervised learning, so in many cases a onetime clustering result of a large data and user desired result will not be the same. Therefore many clustering methods are extended to deal the user constraints. These days, independent components clustering method based on independence of the data distribution have been proposed. In this paper, we propose the method of adding user constraints to the clustering based on the independence of the data distribution. And we show that the proposed method is valid from results of experiments using two artificial datasets.