Ensemble clustering using factor graph
作者：
Highlights：
• Introduce the super-object representation to facilitate the consensus process.
• Probabilistically formulate the ensemble clustering problem into a BLP problem.
• Propose an efficient solver for the BLP problem based on factor graph.
• The cluster number of the consensus clustering is estimated automatically.
• Our method achieves the state-of-the-art performance in effectiveness and efficiency.
摘要
Highlights•Introduce the super-object representation to facilitate the consensus process.•Probabilistically formulate the ensemble clustering problem into a BLP problem.•Propose an efficient solver for the BLP problem based on factor graph.•The cluster number of the consensus clustering is estimated automatically.•Our method achieves the state-of-the-art performance in effectiveness and efficiency.
论文关键词：Ensemble clustering,Factor graph,Belief propagation,Super-object,Automatic cluster number estimate
论文评审过程：Received 24 January 2015, Revised 6 July 2015, Accepted 17 August 2015, Available online 24 August 2015, Version of Record 5 November 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.08.015
