WC-KNNG-PC: Watershed clustering based on k-nearest-neighbor graph and Pauta Criterion

作者：

Highlights：

• Propose a new watershed clustering based on k-nearest-neighbor graph and Pauta Criterion to deal with complex dataset.

• Introduce neighbor information and Pauta Criterion to altitude and point-to-point aggregation.

• propose a two-step immersion to accomplish the point-to-point aggregation.

• This algorithm utilize stability to determine outliers and mergeable subclusters.

• The algorithm presents a basin-level similarity measure to merge sub-clusters.

摘要

•Propose a new watershed clustering based on k-nearest-neighbor graph and Pauta Criterion to deal with complex dataset.•Introduce neighbor information and Pauta Criterion to altitude and point-to-point aggregation.•propose a two-step immersion to accomplish the point-to-point aggregation.•This algorithm utilize stability to determine outliers and mergeable subclusters.•The algorithm presents a basin-level similarity measure to merge sub-clusters.

论文关键词：Watershed clustering,K-nearest neighbor graph (KNNG),Pauta criterion,Shared nearest neighbor (SNN)

论文评审过程：Received 12 October 2020, Revised 9 July 2021, Accepted 14 July 2021, Available online 23 July 2021, Version of Record 1 August 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108177