A reduced variance unsupervised ensemble learning algorithm based on modern portfolio theory
作者：
Highlights：
• This is the first consensus clustering method that takes variance under consideration.
• This is the first application of modern portfolio theory in ensemble learning.
• The algorithm provides reduced variance solutions without a lot of performance sacrifice.
• Algorithm is extensively tested in multiple benchmark instances.
摘要
•This is the first consensus clustering method that takes variance under consideration.•This is the first application of modern portfolio theory in ensemble learning.•The algorithm provides reduced variance solutions without a lot of performance sacrifice.•Algorithm is extensively tested in multiple benchmark instances.
论文关键词：consensus clustering,ensemble learning,internal quality measures,Markowitz’s portfolio theory
论文评审过程：Received 18 March 2020, Revised 5 April 2021, Accepted 17 April 2021, Available online 24 April 2021, Version of Record 8 May 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.115085
