Clustering quality metrics for subspace clustering

作者：

Highlights：

• We present the first study of parameter selection for subspace clustering algorithms.

• A novel point-to-point distance is defined for points on a union of subspaces.

• Several clustering quality metrics specific to subspace clustering are proposed.

• Accounting for underlying geometry in data improves clustering validation results.

摘要

•We present the first study of parameter selection for subspace clustering algorithms.•A novel point-to-point distance is defined for points on a union of subspaces.•Several clustering quality metrics specific to subspace clustering are proposed.•Accounting for underlying geometry in data improves clustering validation results.

论文关键词：Subspace clustering,Clustering validation,Union of subspaces

论文评审过程：Received 22 February 2019, Revised 18 December 2019, Accepted 8 March 2020, Available online 13 March 2020, Version of Record 21 March 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107328