A comparison of AdaBoost algorithms for time series forecast combination
By Devon K. Barrow, Sven F. Crone. 2016
68 citations recorded by the source corpus.
Recently, combination algorithms from machine learning classification have been extended to time series regression, most notably seven variants of the popular AdaBoost algorithm. Despite their theoretical promise their empirical accuracy in forecasting has not yet been assessed, either against each other or against any established approaches of forecast combination, model selection, or statistical benchmark algorithms. Also, none of the algorithms have been assessed on a representative set of empirical data, using only few synthetic time series. We remedy this omission by conducting a rigorous empirical evaluation using a representative set of 111 industry time series and a valid and reliable experimental design. We develop a full-factorial design over derived Boosting meta-parameters, creating 42 novel Boosting variants, and create a further 47 novel Boosting variants using research insights from forecast combination. Experiments show that only few Boosting meta-parameters increase accuracy, while meta-parameters derived from forecast combination research outperform others.
Subfields: Management Science and Operations Research, Signal Processing
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Management Science and Operations Research
- Fotios Petropoulos
- Rob J Hyndman
- Konstantinos Nikolopoulos
- Nikolaos Kourentzes
- Vassilios Assimakopoulos
- Spyros Makridakis
- Evangelos Spiliotis
- Christoph Bergmeir
In Signal Processing
- Vassilios Assimakopoulos
- Konstantinos Nikolopoulos
- Evangelos Spiliotis
- Spyros Makridakis
- Fotios Petropoulos
- Christoph Bergmeir
- Nikolaos Kourentzes
- Devon K. Barrow
Citations in the corpus
- Forecasting: theory and practice (2020, methodology): “1986), while bootstrapping has been used to generate and combine forecast from exponential smoothing (Cordeiro and Neves, 2009; Barrow et al., 2020; Bergmeir et al., 2016), artificial neural networks (Barrow and Crone, 2016a,b), and other forecast methods (Athanasopoulos et al., 2018; Hillebrand and ”
- Forecasting across time series databases using recurrent neural networks on groups of similar series: A clustering approach (2017, background): “Many variations of ensemble architectures have been employed in time series forecasting, such as model stacking (Rahman et al., 2016; Crone et al., 2011), bagging and boosting (Barrow and Crone, 2016; Kourentzes et al., 2014b).”