Proximity-based k-partitions clustering with ranking for document categorization and analysis

作者：

Highlights：

• Weighted medoids are used for cluster representation.

• “Annealing-like” mechanism is introduced to alleviate local optimum.

• Partitions and rankings are obtained by maximizing the objective function.

• Effectiveness and efficiency for document categorization are both reasonable.

摘要

•Weighted medoids are used for cluster representation.•“Annealing-like” mechanism is introduced to alleviate local optimum.•Partitions and rankings are obtained by maximizing the objective function.•Effectiveness and efficiency for document categorization are both reasonable.

论文关键词：Clustering,Similarity-based,k-Medoids,Partitioning,Document categorization

论文评审过程：Available online 17 June 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.06.016