Time series forecasting by neural networks: A knee point-based multiobjective evolutionary algorithm approach

作者：

Highlights：

• Two new multiobjective algorithms NSJADE and KP-NSJADE are proposed.

• The proposed multiobjective algorithms are used to solve the time series forecasting problem.

• Knee point is applied to the time series forecasting problem for the first time.

摘要

•Two new multiobjective algorithms NSJADE and KP-NSJADE are proposed.•The proposed multiobjective algorithms are used to solve the time series forecasting problem.•Knee point is applied to the time series forecasting problem for the first time.

论文关键词：Artificial neural network (ANN),Multiobjective evolutionary algorithm (MOEA),Time series forecasting (TSF),Knee point

论文评审过程：Available online 15 July 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.06.041