Type-1 fuzzy forecasting functions with elastic net regularization

作者：

Highlights：

• A forecasting method based on T1FFs that employs elastic-net is introduced.

• Multicollinearity problem is overcome in the proposed method.

• α and λ are optimized by using nested cross-validation approach.

• Results verified that E-FRF outperformed the selected forecasting benchmarks.

摘要

•A forecasting method based on T1FFs that employs elastic-net is introduced.•Multicollinearity problem is overcome in the proposed method.•α and λ are optimized by using nested cross-validation approach.•Results verified that E-FRF outperformed the selected forecasting benchmarks.

论文关键词：Type-1 fuzzy functions,Elastic-net regularization,Forecasting,Non-linear forecasting

论文评审过程：Received 7 November 2021, Revised 14 February 2022, Accepted 12 March 2022, Available online 21 March 2022, Version of Record 29 March 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.116916