Improved few-shot learning method for transformer fault diagnosis based on approximation space and belief functions
作者：
Highlights：
• Few-shot learning facilitates knowledge extraction from few fault samples.
• Approximate set describes the uncertainty of the diagnostic task in knowledge space.
• Modified belief probability assignment method fosters fault probability estimation.
• Information accumulation promotes accurate fault diagnosis of power transformers.
摘要
•Few-shot learning facilitates knowledge extraction from few fault samples.•Approximate set describes the uncertainty of the diagnostic task in knowledge space.•Modified belief probability assignment method fosters fault probability estimation.•Information accumulation promotes accurate fault diagnosis of power transformers.
论文关键词：Few-shot learning,Rough set,Evidence theory,Approximation space,Belief functions,Transformer fault diagnosis
论文评审过程：Received 1 March 2020, Revised 6 October 2020, Accepted 6 October 2020, Available online 13 October 2020, Version of Record 10 February 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.114105
