Algorithm to determine ε-distance parameter in density based clustering
作者：
Highlights：
• Proposed density based clustering approach adaptively determines ε-distance parameter.
• The methodology is based on the notion of k-nearest neighbours concept.
• Clustering quality of dimensions depend on the data distribution along that dimension.
• Dimensions having clustering quality less than threshold are pruned.
• This makes it appropriate for high dimensional data, as well varying density data.
摘要
•Proposed density based clustering approach adaptively determines ε-distance parameter.•The methodology is based on the notion of k-nearest neighbours concept.•Clustering quality of dimensions depend on the data distribution along that dimension.•Dimensions having clustering quality less than threshold are pruned.•This makes it appropriate for high dimensional data, as well varying density data.
论文关键词：Data mining,Clustering,Density based clustering,Subspace clustering,High dimensional data
论文评审过程：Available online 30 October 2013.
论文官网地址：https://doi.org/10.1016/j.eswa.2013.10.025
