Relative performance of methods for forecasting special events

Nikolopoulos, Konstantinos and Litsa, Akrivi and Petropoulos, Fotios and Bougioukos, Vasileios and Khammash, Marwan (2015) Relative performance of methods for forecasting special events. Journal of Business Research, 68 (8). pp. 1785-1791. ISSN 0148-2963

Abstract

Forecasting special events such as conflicts and epidemics can be very challenging, not only due to the nature of those events but also for the limited amount of historical information from which a reference base can be built. The present study evaluates structured analogies, Delphi and interaction groups for forecasting the impact of such events. Empirical evidence reveals that when using structured analogies an average accuracy improvement of 5.1% is achieved compared to unaided judgment. This improvement is greater when: a) the level of expertise increases, b) more analogies are used, c) more-relevant analogies are used, and d) pooling of analogies is introduced through interaction of the experts. Furthermore, group judgmental forecasting approaches illustrated promising results with Delphi group having an improved accuracy of 22.3% and Interaction Group of 40.6% respectively.