Monday, January 4, 2016

GEFCom2014 Probabilistic Electric Load Forecasting: An Integrated Solution with Forecast Combination and Residual Simulation

Guest Blogger: Jingrui Xie

My adventure on load forecasting started in 2012, when I was the primary developer for the SAS Energy Forecasting solution. Our first customer was North Carolina Electric Membership Cooperation, who used the probabilistic load forecasts generated from our solution for long-term power supply planning.  That project was documented in my first TSG paper co-authored with Dr. Tao Hong and Jason Wilson, which was titled Long term probabilistic load forecasting and normalization with hourly information. Later on, the method proposed in that TSG paper was further investigated by analyzing the residuals. The findings were summarized in our recently published TSG paper On normality assumption in residual simulation for probabilistic load forecasting.

With all these experiences in load forecasting, I was very excited to join GEFCom2014. I thought it would be a great opportunity for me to share my probabilistic load forecasting experiences with the worldwide load forecasting community. Right after the competition started, I delivered my daughter. I tried to squeeze some time to work on this competition after putting my newborn daughter to sleep. After the 15-week long competition, I was ranked top 3 in the final leaderboard. Many thanks to my husband who supported me through this.

Today, I am very pleased to announce that the paper summarizing the methodologies I took for the probabilistic electric load forecasting track of GEFCom2014 has just been published by International Journal of Forecasting. In addition to the methodologies written in the two TSG papers mentioned above, this IJF paper also includes combining forecasts from various statistical and artificial intelligence techniques. The forecast combination component is largely similar to the SAS Energy Forecasting solution I developed with Dr. Hong and the SAS colleagues in early 2012.

Finally, I would like to give my thanks to Dr. Tao Hong, Dr. Shu Fan and several other anonymous reviewers for their valuable comments. I hope this paper is helpful to the load forecasting community.


Jingrui Xie, Tao Hong, GEFCom2014 probabilistic electric load forecasting: An integrated solution with forecast combination and residual simulation, International Journal of Forecasting, Available online 19 December 2015, ISSN 0169-2070,

GEFCom2014 Probabilistic Electric Load Forecasting: An Integrated Solution with Forecast Combination and Residual Simulation

Jingrui Xie and Tao Hong

We present an integrated solution for probabilistic load forecasting. The proposed solution was the basis for Jingrui Xie’s submission to the probabilistic load forecasting track of the Global Energy Forecasting Competition 2014 (GEFCom2014), and consists of three components: pre-processing, forecasting, and post-processing. The pre-processing component includes data cleansing and temperature station selection. The forecasting component involves the development of point forecasting models, forecast combination, and temperature scenario based probabilistic forecasting. The post-processing component embodies residual simulation for probabilistic forecasting. In addition, we also discuss several other variations that were implemented during the competition.


  1. Dear Jingrui Xie,

    Compliments for your work!

    I was wondering if one of the participants used your method in the last in-class probabilistic load forecasting competition organized by Tao Hong.

    Best regards, Geert

  2. Jingrui's method used in the competition was quite comprehensive. None of the students in that in-class competition used the method as comprehensive as Jingrui's. For instance, none of them used forecast combination. Most of them used a simplified version of the method in our TSG paper published in 2014.


Note that you may link to your LinkedIn profile if you choose Name/URL option.