Text Clustering via Constrained Nonnegative Matrix Factorization

摘要: Semi-supervised nonnegative matrix factorization (NMF)receives more and attention in text mining field. The semi-supervised NMF methods can be divided into two types, one is based on the explicit category labels, other pair wise constraints including must-link cannot-link. As it hard to obtain labels some tasks, latter widely used real applications. To date, all constrained treat cannot-link a same way. However, these kinds of play different roles clustering. Thus novel method proposed this paper. In new method, are control distance data compressed form, cannot-ink encoding factor. Experimental results real-world sets have shown good performance method.