Information theoretic clustering using a k-nearest neighbors approach
作者：
Highlights：
• Propose a new method of estimating Information Theoretic measures using KNN.
• Introduce a hierarchical clustering routine using this estimate.
• Use two different values for k depending on which information theoretic measure is being estimated.
• Avoid having to tune a critical parameter for each clustering task.
• Handles datasets of different scales well compared to traditional methods.
摘要
Highlights•Propose a new method of estimating Information Theoretic measures using KNN.•Introduce a hierarchical clustering routine using this estimate.•Use two different values for k depending on which information theoretic measure is being estimated.•Avoid having to tune a critical parameter for each clustering task.•Handles datasets of different scales well compared to traditional methods.
论文关键词：Clustering,Scale,Entropy,Divergence,k-nn,Parzen windowing,Information theory
论文评审过程：Received 19 March 2013, Revised 17 February 2014, Accepted 18 March 2014, Available online 28 March 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.03.018
