K-means⁎: Clustering by gradual data transformation

作者：

Highlights：

• Traditionally clustering is done by fitting the clustering model to the data.

• We propose an opposite approach by fitting the data into a given clustering model.

• We perform inverse transform from this pathological data back to the original data.

• We refine the optimal clustering structure during the process.

摘要

Highlights•Traditionally clustering is done by fitting the clustering model to the data.•We propose an opposite approach by fitting the data into a given clustering model.•We perform inverse transform from this pathological data back to the original data.•We refine the optimal clustering structure during the process.

论文关键词：Clustering,K-means,Data transformation

论文评审过程：Received 30 September 2013, Revised 27 March 2014, Accepted 29 March 2014, Available online 18 April 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.03.034