Using a rolling training approach to improve judgmental extrapolations elicited from forecasters with technical knowledge

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Several biases and inefficiencies are commonly associated with the judgmental extrapolation of time series even when forecasters have technical knowledge about forecasting. This study examines the effectiveness of using a rolling training approach, based on feedback, to improve the accuracy of forecasts elicited from people with such knowledge. In an experiment forecasters were asked to make multiple judgmental extrapolations for a set of time series from different time origins. For each series in turn, the participants were either unaided or they were provided with feedback. In the latter case, following submission of each set of forecasts, the true outcomes and performance feedback were provided. The objective was to provide a training scheme, enabling forecasters to better understand the underlying pattern of the data by learning directly from their forecast errors. Analysis of the results indicated that the rolling training approach is an effective method for enhancing judgmental extrapolations elicited from people with technical knowledge, especially when bias feedback is provided. As such it can be a valuable element in the design of software systems that are intended to support expert knowledge elicitation (EKE) in forecasting.
LanguageEnglish
Pages314-324
Number of pages11
JournalInternational Journal of Forecasting
Volume33
Issue number1
Early online date30 Mar 2016
DOIs
StatusPublished - Jan 2017

Fingerprint

Extrapolation
Forecast error
Performance feedback
Inefficiency
Experiment
Systems software
Knowledge elicitation
Expert knowledge

Cite this

@article{7c08bfe238e44c1388d72f140cc0a883,
title = "Using a rolling training approach to improve judgmental extrapolations elicited from forecasters with technical knowledge",
abstract = "Several biases and inefficiencies are commonly associated with the judgmental extrapolation of time series even when forecasters have technical knowledge about forecasting. This study examines the effectiveness of using a rolling training approach, based on feedback, to improve the accuracy of forecasts elicited from people with such knowledge. In an experiment forecasters were asked to make multiple judgmental extrapolations for a set of time series from different time origins. For each series in turn, the participants were either unaided or they were provided with feedback. In the latter case, following submission of each set of forecasts, the true outcomes and performance feedback were provided. The objective was to provide a training scheme, enabling forecasters to better understand the underlying pattern of the data by learning directly from their forecast errors. Analysis of the results indicated that the rolling training approach is an effective method for enhancing judgmental extrapolations elicited from people with technical knowledge, especially when bias feedback is provided. As such it can be a valuable element in the design of software systems that are intended to support expert knowledge elicitation (EKE) in forecasting.",
author = "Fotios Petropoulos and Paul Goodwin and Robert Fildes",
year = "2017",
month = "1",
doi = "10.1016/j.ijforecast.2015.12.006",
language = "English",
volume = "33",
pages = "314--324",
journal = "International Journal of Forecasting",
issn = "0169-2070",
publisher = "Elsevier",
number = "1",

}

TY - JOUR

T1 - Using a rolling training approach to improve judgmental extrapolations elicited from forecasters with technical knowledge

AU - Petropoulos, Fotios

AU - Goodwin, Paul

AU - Fildes, Robert

PY - 2017/1

Y1 - 2017/1

N2 - Several biases and inefficiencies are commonly associated with the judgmental extrapolation of time series even when forecasters have technical knowledge about forecasting. This study examines the effectiveness of using a rolling training approach, based on feedback, to improve the accuracy of forecasts elicited from people with such knowledge. In an experiment forecasters were asked to make multiple judgmental extrapolations for a set of time series from different time origins. For each series in turn, the participants were either unaided or they were provided with feedback. In the latter case, following submission of each set of forecasts, the true outcomes and performance feedback were provided. The objective was to provide a training scheme, enabling forecasters to better understand the underlying pattern of the data by learning directly from their forecast errors. Analysis of the results indicated that the rolling training approach is an effective method for enhancing judgmental extrapolations elicited from people with technical knowledge, especially when bias feedback is provided. As such it can be a valuable element in the design of software systems that are intended to support expert knowledge elicitation (EKE) in forecasting.

AB - Several biases and inefficiencies are commonly associated with the judgmental extrapolation of time series even when forecasters have technical knowledge about forecasting. This study examines the effectiveness of using a rolling training approach, based on feedback, to improve the accuracy of forecasts elicited from people with such knowledge. In an experiment forecasters were asked to make multiple judgmental extrapolations for a set of time series from different time origins. For each series in turn, the participants were either unaided or they were provided with feedback. In the latter case, following submission of each set of forecasts, the true outcomes and performance feedback were provided. The objective was to provide a training scheme, enabling forecasters to better understand the underlying pattern of the data by learning directly from their forecast errors. Analysis of the results indicated that the rolling training approach is an effective method for enhancing judgmental extrapolations elicited from people with technical knowledge, especially when bias feedback is provided. As such it can be a valuable element in the design of software systems that are intended to support expert knowledge elicitation (EKE) in forecasting.

UR - http://dx.doi.org/10.1016/j.ijforecast.2015.12.006

UR - http://dx.doi.org/10.1016/j.ijforecast.2015.12.006

U2 - 10.1016/j.ijforecast.2015.12.006

DO - 10.1016/j.ijforecast.2015.12.006

M3 - Article

VL - 33

SP - 314

EP - 324

JO - International Journal of Forecasting

T2 - International Journal of Forecasting

JF - International Journal of Forecasting

SN - 0169-2070

IS - 1

ER -