A generous donation from Professor Tao Hong has funded this award for papers on energy forecasting published in the International Journal of Forecasting. Every two years the IJF editors will select the “best” paper on energy forecasting to have been published in the IJF within the previous two-year period. The IIF Tao Hong Award consists of US$1000 and an engraved plaque. Additional papers may receive Outstanding Paper Awards.
2019-2020 IIF Tao Hong Award
Peru Muniain and Florian Ziel (2020) “Probabilistic forecasting in day-ahead electricity markets: Simulating peak and off-peak prices”, IJF 36(4), 1193-1210.
The paper by Muniain and Ziel looks at a problem that has already been the focus of many previous works (i.e., price forecasting in day-ahead electricity markets), but with a fresh twist to it. Indeed, the authors first push towards probabilistic forecasting, but in a multivariate environment. Indeed, considering the timing of peak and off-peak prices requires to think about the temporal dependence structure in the probabilistic forecasts. The nominators, as well as the award committee, underline the quality of the work, the fact that this is one of the first papers to look at multivariate probabilistic forecast verification within electricity price forecasting, while expecting that the paper may make a strong impact in the coming years.
2019-2020 Outstanding Paper Awards
Bartosz Uniejewski, Grzegorz Marcjasz and Rafał Weron (2019) “Understanding intraday electricity markets: Variable selection and very short-term price forecasting using LASSO”, IJF 35(4), 1533-1547.
This is a very detailed piece of work with focus on the intra-day electricity prices in the EPEX in Germany. The work is very extensive with a strong focus on variable selection and analysing the way these variables may or may not be relevant depending on certain regimes, time of week, etc. Interestingly also, the paper aims at linking forecast quality and value by looking at the resulting benefits from using the forecasts as input to simple market trading strategies.
Jorge Ángel González Ordiano, Lutz Gröll, Ralf Mikut, Veit Hagenmeyer (2020) “Probabilistic energy forecasting using the nearest neighbors quantile filter and quantile regression”, IJF 36(2), 310-323.
The authors develop a nice and practical nonparametric approach to probabilistic forecasting, which they refer to as nearest neighbors quantile filter (NNQF), with an application to solar energy forecasting (GEFCom 2014 dataset) to illustrate its workings and performance. The main argument is that the approach is very easy to use, very cheap computationally, while being very close in terms of performance to the best entries of the GEFcom 2014 competition.
2017-2018 IIF Tao Hong Award
Kevin Berk, Alexander Hoffmann and Alfred Müller (2018) “Probabilistic forecasting of industrial electricity load with regime-switching behaviour”, IJF 34(2), 147-162.
The paper by Berk, Hoffmann and Müller concentrates on an important problem in energy forecasting, related to electricity demand. While a majority of the electric load forecasting works, until a few years ago, dealt with electric load prediction at an aggregate level, the deployment of smart meters and new challenges in electric energy management (renewables, demand response, liberalization, etc.) makes that it is of utmost importance to now focus on also predicting load at very detailed levels, e.g. for households and industry. This is what the paper delivers, by rightly appraising the underlying characteristics of industrial load and their potential regime-switching nature. The paper elegantly blends approaches time-series modelling, regime-switching (with time-varying transition probabilities) and a dedicated estimation. Eventually, the forecasts generated are probabilistic in nature and evaluated as such.
2015-2016 IIF Tao Hong Award
Pierre Gaillard, Yannig Goude and Raphael Nedellec (2016) “Additive models and robust aggregation for GEFCom2014 probabilistic electric load and electricity price forecasting”, IJF 32(3), 1038-1050.
This EDF R&D team won the load and price forecasting tracks of the GEFCom2014 competition. The load forecasting track, in particular, was very competitive. The paper touches three different methods, one used for both tracks (quantile generalized additive model), a combination method, and a sparse regression based method. All of these three approaches are novel and can be considered as the state of the art today.
2015-2016 Outstanding Paper Awards
Jaromir Benes, Marcelle Chauvet, Ondra Kamenik, Michael Kumhof, Douglas Laxton, Susanna Mursula, Jack Selody (2015) “The future of oil: Geology versus technology”, IJF 31(1), 207-221.
This paper takes a general perspective on how to forecast oil prices and oil output that nests both economic and geologic views. The authors put their model to various uses, focusing on the forecasting, but also using their model to interpret what type of shocks explain oil price hikes and to explore the links between oil and the macro-economy.
2013-2014 IIF Tao Hong Award
Rafał Weron (2014) “Electricity price forecasting: A review of the state-of-the-art with a look into the future”, IJF 30(4), 1030–1081.
Weron’s article is encyclopedic. It covers the entirety of electricity price forecasting and systematizes a wide field of very disparate models, from statistical models across machine learning to agent models. However, the paper is not only a review. It also presents much needed guidelines for the rigorous use of methods, measures and tests, and it looks ahead and speculates on the directions electricity price forecasting will take in the next decade or so. The paper also has applications beyond electricity price forecasting, and is of particular interest to people in electric load forecasting and call center forecasting.
2013-2014 Outstanding Paper Awards
Stephen Haben, Jonathan Ward, Danica Vukadinovic Greetham, Colin Singleton and Peter Grindrod (2014) “A new error measure for forecasts of household-level, high resolution electrical energy consumption”, IJF 30(2), 246–256.
Haben et al (2014) provides high-quality verification tools for load forecasts, which are essential in managing power systems. The paper is particularly helpful for work on demand profiling in the residential sector, where the temporal resolution of data has increased rapidly in recent years. An accurate and ‘fair’ verification method for these highly volatile data is of great importance. In addition, the paper is sufficiently accessible that it can be easily applied in practice.