Meta‐learning how to forecast time series
Journal of Forecasting · 2023-09 · 42 (6) · 1476-1501
DOI: 10.1002/for.2963 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Forecasting, online erschienen am 24.02.2023. Verfasst von Thiyanga S. Talagala und 2 weiteren Autor:innen. Die Arbeit wurde bisher 50-mal zitiert (Crossref) und nennt 54 Literaturangaben.
Abstract (Originalsprache)
Features of time series are useful in identifying suitable models for forecasting. We present a general framework, labelled Feature‐based FORecast Model Selection (FFORMS), which selects forecast models based on features calculated from each time series. The FFORMS framework builds a mapping that relates the features of a time series to the “best” forecast model using a classification algorithm such as a random forest. The framework is evaluated using time series from the M‐forecasting competitions and is shown to yield forecasts that are almost as accurate as state‐of‐the‐art methods but are much faster to compute. We use model‐agnostic machine learning interpretability methods to explore the results and to study what types of time series are best suited to each forecasting model.
Zitieren
APA
Talagala, T. S., Hyndman, R. J., & Athanasopoulos, G. (2023). Meta‐learning how to forecast time series. Journal of Forecasting, 42(6), 1476-1501. https://doi.org/10.1002/for.2963
Vancouver
Talagala TS, Hyndman RJ, Athanasopoulos G. Meta‐learning how to forecast time series. Journal of Forecasting. 2023;42(6):1476-1501. doi:10.1002/for.2963
BibTeX
@article{Talagala2023,
author = {Talagala, Thiyanga S. and Hyndman, Rob J. and Athanasopoulos, George},
title = {Meta‐learning how to forecast time series},
journal = {Journal of Forecasting},
year = {2023},
volume = {42},
number = {6},
pages = {1476--1501},
doi = {10.1002/for.2963}
}
EUVIMED