To cluster, or not to cluster: An analysis of clusterability methods

作者：

Highlights：

• The paper surveys and compares clusterability tests.

• New clusterability tests are proposed.

• Type I error and power of clusterability methods are reported for simulated data.

• Clusterablity tests are applied to well-known non-simulated data.

• Provide guidelines to help users to select among clusterability tests.

摘要

•The paper surveys and compares clusterability tests.•New clusterability tests are proposed.•Type I error and power of clusterability methods are reported for simulated data.•Clusterablity tests are applied to well-known non-simulated data.•Provide guidelines to help users to select among clusterability tests.

论文关键词：Clusterability,Cluster structure,Cluster tendency,Dimension reduction,Multimodality tests

论文评审过程：Received 2 September 2017, Revised 7 February 2018, Accepted 21 October 2018, Available online 25 October 2018, Version of Record 9 November 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.10.026