Diversity-driven generation of link-based cluster ensemble and application to data classification

作者：

Highlights：

• New data-transformation method that makes use of link-based cluster ensemble (LCE).

• For accurate clustering, LCE is coupled with diversity-driven ensemble generation.

• Evaluated on published datasets with C4.5, NB, KNN, ANN and Random Forest models.

• New method usually performs better than benckmark techniques.

摘要

Highlights•New data-transformation method that makes use of link-based cluster ensemble (LCE).•For accurate clustering, LCE is coupled with diversity-driven ensemble generation.•Evaluated on published datasets with C4.5, NB, KNN, ANN and Random Forest models.•New method usually performs better than benckmark techniques.

论文关键词：Ensemble clustering,Data classification,Optimization,Feature transformation

论文评审过程：Available online 18 July 2015, Version of Record 25 July 2015.

论文官网地址：https://doi.org/10.1016/j.eswa.2015.06.051