Tuning Forecasting Algorithms for Black Swan Events
Publication Type:
Refereed Conference Meeting Proceeding
Abstract:
Forecasting algorithms based on exponential smoothing have smoothing
factors, and it is often recommended that these be tuned to minimise
an error measure on observed data. We show that forecasting
algorithms such as simple exponential smoothing and Croston's method
cannot always be optimally tuned to time series using any of several
error measures. We propose a data augmentation approach: adding
hypothetical non-stationary time series (which we call ``black swans''
as they represent unseen pathological cases) to the training data, and
minimising a weighted error. The choice of black swans is a form of
judgemental forecasting that requires experts to explicitly state
their assumptions on unseen data.
Conference Name:
9th IFAC Conference on Manufacturing Modelling, Management and Control
Digital Object Identifer (DOI):
10.1016/j.ifacol.2019.11.411
Publication Date:
20/02/2019
Volume:
53
Issue:
13
Pages:
1496-1501
Conference Location:
Germany
Research Group:
Institution:
National University of Ireland, Cork (UCC)
Open access repository:
No
Publication document: