摘要: Constrained clustering received a lot of attention in the last years. However, widely used pairwise constraints are not generally applicable for hierarchical clustering, where goal is to derive cluster hierarchy instead flat partition. Therefore, we propose setting--based on ideas constraints--the use must-link-before (MLB) constraints. In this paper, discuss their properties and present an algorithm that able create by considering these directly. Furthermore, efficient data structure its implementation evaluate effectiveness with different datasets text scenario.