基于成对约束的半监督凝聚层次聚类算法

对于所提出的建立在成对约束基础之上的半监督凝聚层次聚类算法,对聚类簇进行半监督处理的最主要目的在于借助于对样本监督信息的合理应用,达到提高样本在无监督状态下学习性能的目标.在现阶段的技术条件支持下,以半监督聚类分析为核心,建立在must link以及cannot link基础之上的约束关系被广泛地应用于样本聚类分析的过程当中.从这一角度上来说,为了使聚类簇与聚类簇之间的距离关系表述更加的真实与精确,就要求通过对成对约束关系的综合应用,实现对聚类簇距离的有效调整与优化.

Semi-supervised Agglomerative Hierarchical Clustering Algorithm Based on Pairwise Constraints

In this paper, semi-supervised agglomerative hierarchical clustering algorithm is based on pairwise constraints. The main purpose of semi-supervised processing Clustering in the absence of state supervision, the rational use of sample information can improve the sample learning performance. With the support of modem technical condition, based on semi- supervised clustering, the constraint relationship between must link and cannot link is established, the constraint relations is widely used in the cluster analysis of sample. So, from this point, in order to make the relationship between two clusters more real and accurate, comprehensive application of pairwise constraints relationships can adjust and optimize the distance between two clusters effectively.