A heuristic approach to effective and efficient clustering on uncertain objects

作者：

Highlights：

• We reduce UK-means to K-means.

• We experimentally show that K-means performs much faster than existing pruning algorithms.

• We propose Approximate UK-means to heuristically identify boundary objects and re-assign them to better clusters.

• We propose three models for the representation of cluster representative. To our knowledge, this is the first time to introduce uncertain model of cluster representative.

摘要

•We reduce UK-means to K-means.•We experimentally show that K-means performs much faster than existing pruning algorithms.•We propose Approximate UK-means to heuristically identify boundary objects and re-assign them to better clusters.•We propose three models for the representation of cluster representative. To our knowledge, this is the first time to introduce uncertain model of cluster representative.

论文关键词：Clustering,Uncertain objects,UK-means,Expected Euclidean distance,Expected squared Euclidean distance

论文评审过程：Received 1 September 2013, Revised 3 April 2014, Accepted 17 April 2014, Available online 24 April 2014.

论文官网地址：https://doi.org/10.1016/j.knosys.2014.04.027