Global Energy Forecasting Competition (Oracle, Mar14)

Energy forecasting, in a broad sense, covers a wide range of forecasting problems in the energy industry, such as demand forecasting, generation forecasting, price forecasting and so on. These problems are very attractive to the forecasting community due to the characteristics including high resolution data, multiple seasonality, requirements of low errors and societal necessity.

Despite more than 100 years of practice (see the recent Foresight article – Energy Forecasting: Past, Present and Future), there are still many major issues with the field of energy forecasting, such as lack of benchmark data and processes, irreproducible methodologies and results from most papers in the literature, and the large gap between industry needs and academic education and research activities. Given these facts, we organized the Global Energy Forecasting Competition 2012 (GEFCom2012) in order to (i) improve the forecasting practices of the utility industry, (ii) bring together state-of-the-art techniques for energy forecasting, (iii) bridge the gap between academic research and industry practice, (iv) promote analytics in power and energy education, and (v) prepare the industry to overcome the forecasting challenges brought by the smart grid technologies and renewable integration needs.

GEFCom2012, the largest energy forecasting competition known to date, attracted hundreds of data scientists from over 30 countries. Over the two-month competition period, more than 200 teams submitted more than 2,000 entries to the two tracks of GEFCom2012, hierarchical load forecasting and wind power forecasting. The 8 winning teams were formulated by people from 8 countries. The success of GEFCom2012 was reported by major media channels and professional organizations, such as CNBC, Yahoo Finance, Reuters, IEEE Power and Energy Society and Forecasting Principles.

The next competition GEFCom2014 (www.gefcom.org) will feature three major upgrades: 1) probabilistic forecasts in the form of predicted quantiles; 2) four tracks on demand, price, wind and solar; 3) rolling forecasts with incremental data update on weekly basis. We will launch the competition on August 15th and end it on December 15th of 2014. GEFCom2014 will recognize up to 3 winning teams per track. For the first time, we will recognize the top 3 academic institutes with multiple leading teams and offer a prize for the top one institute (see Announcing Institute Prize for GEFCom2014). The International Journal of Forecasting will publish a special issue on Probabilistic Energy Forecasting, which will archive the data and winning methodologies of GEFCom2014 in addition to the original research papers under the same theme (see IJF Call For Papers: Probabilistic Energy Forecasting).

As the General Chair of the Global Energy Forecasting Competition, I was fortunate to have strong support from Drs. Pierre Pinson (Technical University of Denmark, Denmark), Shu Fan and Rob Hyndman (both from Monash University, Australia) and many other individuals and organizations to make GEFCom2012 a fantastic event. This year, we have two more experts joining the executive committee, Drs. Hamidreza Zareipour (University of Calgary, Canada) and Alberto Troccoli (CSIRO, Australia). I’m sure that we are going to present you the modern challenges of energy forecasting.

Are you ready to show off your forecasting skills? Visit www.gefcom.org to join the growing interest list with 130+ forecasters from 39 countries and SARs.

Dr. Tao Hong is Graduate Program Director and EPIC Assistant Professor of Systems Engineering and Engineering Management at University of North Carolina at Charlotte. He is the Founding Chair of IEEE Working Group on Energy Forecasting, General Chair of Global Energy Forecasting Competition and lead author of the online textbook Electric Load Forecasting: Fundamentals and Best Practices.

Share This: