A new methodology for generating and combining statistical forecasting models to enhance competitive event prediction

Lessmann, Stefan; Sung, Ming‐Chien; Johnson, J.E.V.; Ma, Tiejun

doi:10.1016/j.ejor.2011.10.032

Cited by 23 publications

(24 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most common combination forecasting models are linear combination of all available individual forecast models in tourism literature. The researchers (Andrawisa et al, 2011, Chan et al, 2010Coshall & Charlesworth, 2011;Freitas & Rodrigues, 2006;Lessmann et al, 2012;Menezes, Bunn & Taylor, 2000;Shen, Li & Song, 2011) have demonstrated the efficiency of combination forecasts and the superiority of combination forecasts in contrast to individual forecasts. However all available individual models are used as inputs for the combinations.…”

Section: Introductionmentioning

confidence: 99%

“…Since then it has been demonstrated that forecast combinations are often superior to their constituent forecasts in many fields (Greer, 2005;Hall & Mitchell, 2007;Holden & Peel, 1986;Lessmann et al (2012) ;Li, Shi & Zhou, 2011;Newbold & Granger, 1974;Sánchez, 2 2008;Timmermann, Elliott & Granger, 2006;Winkler & Makridakis, 1983;Zheng, Lee & Shi, 2006). The most widely used and studied combination forecast methods are ensemble methods, such as bagging (Breiman, 1996) and boosting methods.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A combination selection algorithm on forecasting

Cang

2014

European Journal of Operational Research

View full text Add to dashboard Cite

It is widely accepted in forecasting that a combination model can improve forecasting accuracy. One important challenge is how to select the optimal subset of individual models from all available models without having to try all possible combinations of these models. This paper proposes an optimal subset selection algorithm from all individual models using information theory. The experimental results in tourism demand forecasting demonstrate that the combination of the individual models from the selected optimal subset significantly outperforms the combination of all available individual models. The proposed optimal subset selection algorithm provides a theoretical approach rather than experimental assessments which dominate literature.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A combination selection algorithm on forecasting

Cang

2014

European Journal of Operational Research

View full text Add to dashboard Cite

show abstract

“…In such settings, pseudo-R 2 s, which are equivalent measures to R 2 for nonlinear models, are generally recommended (e.g., Greene, 2012;Maddala, 1983). In fact, in any environment where one seeks to use probability predictions to make pecuniary gain (e.g., in options, futures, spread trading and betting markets), it can be shown that there is a direct link between increases in pseudo-R 2 values and out-of-sample returns (e.g., Benter, 1994;Lessmann et al, 2012).…”

Section: Introductionmentioning

confidence: 99%

“…One of the many applications for discrete choice models is in the forecasting 4 of competitive event (CE) outcomes (e.g., Lessmann et al, 2009Lessmann et al, , 2012. A CE is a contest between at least two rival participants, where (generally) one winner is declared and the outcome is uncertain, such as political elections or sporting events.…”

Section: Introductionmentioning

confidence: 99%

Probabilistic forecasting with discrete choice models: Evaluating predictions with pseudo-coefficients of determination

Sung

McDonald

Johnson

2016

European Journal of Operational Research

Self Cite

View full text Add to dashboard Cite

Probabilistic forecasts from discrete choice models, which are widely used in marketing science and competitive event forecasting, are often best evaluated out-of-sample using pseudo-coefficients of determination, or pseudo-R 2 s. However, there is a danger of misjudging the accuracy of forecast probabilities of event outcomes, based on observed frequencies, because of issues related to pseudo-R 2 s. First, we show that McFadden's pseudo-R 2 varies predictably with the number of alternatives in the choice set. Then we evaluate the relative merits of two methods (bootstrap and asymptotic) for estimating the variance of pseudo-R 2 s so that their values can be appropriately compared across nonnested models. Finally, in the context of competitive event forecasting, where the accuracy of forecasts has direct economic consequence, we derive new R 2 measures that can be used to assess the economic value of forecasts. Throughout, we illustrate using data drawn from UK and Ireland horse race betting markets.

show abstract

“…Forecasts still remain far from perfect, and may fall short of reaching society's expectations for timely and reliable warnings. Therefore, the need arises to improve prediction system by, for example, through combing physical and empirical models (Goetz et al, 2011;Ziervogel et al, 2005;Lessmann et al, 2012). Several studies of empirical modelling have been carried out to predict stream-flow (OpitzStapleton et al, 2007;Barnston et al, 1999;Tisseuil et al, 2010;Krzysztofowicz., 1999;Goetz et al, 2011;Maier et al, 2000); however most of these approaches still have short-comings to skilfully predict the flows.…”

Section: Introductionmentioning

confidence: 99%

Prediction of inflows into Lake Kariba using a combination of physical and empirical models

Muchuru

Landman

DeWitt

2015

Intl Journal of Climatology

View full text Add to dashboard Cite

Seasonal climate forecasts are operationally produced at various climate prediction centres around the world. However, these forecasts may not necessarily be objectively integrated into application models in order to help with decision-making processes. The use of hydro-meteorological models may be proven effective for reservoir operations since accurate and reliable prediction of reservoir inflows can provide balanced solution to the problems faced by dam or reservoir managers. This study investigates the use of a combination of physical and empirical models to predict seasonal inflows into Lake Kariba in southern Africa. Two predictions systems are considered. The first uses antecedent seasonal rainfall totals over the upper Zambezi catchment as predictor in a statistical model for estimating seasonal inflows into Lake Kariba. The second and more sophisticated method uses predicted low-level atmospheric circulation of a coupled ocean-atmosphere general circulation model (CGCM) downscaled to the inflows. Forecast verification results are presented for five run-on 3-month seasons; from September to June over an independent hindcast period of 14 years (1995/6 to 2008/9). Verification is conducted using the relative operating characteristic (ROC) and the reliability diagram. In addition to the presented verification statistics, the hindcasts are also evaluated in terms of their economic value as a usefulness indicator of forecast quality for bureaucrats and to the general public. The models in general perform best during the austral mid-summer season of DJF (seasonal onset of inflows) and the autumn season of MAM (main inflow season). Moreover, the prediction system that uses the output of the CGCM is superior to the simple statistical approach. An additional forecast of a recent flooding event (2010/11), which lies outside of the 14-year verification window, is presented to further demonstrate the forecast system's operational capability during a season of high inflows that caused societal and infrastructure problems over the region.

show abstract

A new methodology for generating and combining statistical forecasting models to enhance competitive event prediction

Cited by 23 publications

References 60 publications

A combination selection algorithm on forecasting

A combination selection algorithm on forecasting

Probabilistic forecasting with discrete choice models: Evaluating predictions with pseudo-coefficients of determination

Prediction of inflows into Lake Kariba using a combination of physical and empirical models

Contact Info

Product

Resources

About