Modified fuzzy regression functions with a noise cluster against outlier contamination

作者：

Highlights：

• A new modified fuzzy regression functions framework is proposed.

• The new framework is robust against outlier contamination.

• Possibilistic and fuzzy-possibilistic clustering is employed.

• The proposed framework performs better than ANNs and SVMs.

• Modified fuzzy-possibilistic clustering is a robust algorithm against outliers.

摘要

•A new modified fuzzy regression functions framework is proposed.•The new framework is robust against outlier contamination.•Possibilistic and fuzzy-possibilistic clustering is employed.•The proposed framework performs better than ANNs and SVMs.•Modified fuzzy-possibilistic clustering is a robust algorithm against outliers.

论文关键词：Artificial neural networks,Noise cluster,Outlier,Possibilistic clustering,Robustness,Support vector machines

论文评审过程：Received 17 March 2021, Revised 7 April 2022, Accepted 30 May 2022, Available online 2 June 2022, Version of Record 6 June 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117717