An experimental methodology to evaluate machine learning methods for fault diagnosis based on vibration signals

作者：

Highlights：

• Systematic analysis of overoptimistic results in machine learning fault diagnosis.

• Computational framework to test new feature models.

• Computational framework to test new classifier architectures

• Experimental analysis of more realistic fault diagnosis scenarios.

摘要

•Systematic analysis of overoptimistic results in machine learning fault diagnosis.•Computational framework to test new feature models.•Computational framework to test new classifier architectures•Experimental analysis of more realistic fault diagnosis scenarios.

论文关键词：Fault detection,CWRU bearing fault database,Performance criteria,Classification,Pattern recognition,Machine learning

论文评审过程：Received 6 April 2020, Revised 22 August 2020, Accepted 14 September 2020, Available online 30 September 2020, Version of Record 10 February 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.114022