A multi-method forecasting algorithm: Linear unbiased estimation of combine forecast
作者：
Highlights：
• Best possible combinations of forecasters to have better predictive performance.
• Sparse weighting algorithm to combine forecasts to reduce forecasting error.
• Linear combination of forecaster vs average and median combining methods.
• Dynamic ensemble algorithm to combine forecasting results from multiple methodologies.
• Self-calibrating forecasting algorithm combining various models.
摘要
•Best possible combinations of forecasters to have better predictive performance.•Sparse weighting algorithm to combine forecasts to reduce forecasting error.•Linear combination of forecaster vs average and median combining methods.•Dynamic ensemble algorithm to combine forecasting results from multiple methodologies.•Self-calibrating forecasting algorithm combining various models.
论文关键词：Combine forecasting,Model learning,Sparse model combination
论文评审过程：Received 7 October 2021, Revised 3 December 2021, Accepted 15 December 2021, Available online 21 December 2021, Version of Record 1 January 2022.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.107990
