Improving Forecast Accuracy Using a Synthetic Weather Station: An Incremental Approach and BFCom2018 Lessons Leamed

Selection of the most appropriate forecast model should be governed by the underlying data. This paper investigates the impact of benchmark model selection, recency effect and synthetic weather station selection techniques on load forecast performance and presents a new weighted average based approach to generate a synthetic weather station. Lessons learned from this effort include the criticality of using benchmark models, the need for additional public datasets, and the value of forecasting competitions for learning and model development. The results from this case study validate that addition of recency effect and use of a synthetic weather station, can provide substantial forecast improvements over benchmark models. The results also demonstrate the potential benefit of using a weighted approach for synthetic weather station generation rather than simple averaging.

[1]  Rob J Hyndman,et al.  Forecasting Time Series With Complex Seasonal Patterns Using Exponential Smoothing , 2011 .

[2]  T. Hesterberg,et al.  A regression-based approach to short-term system load forecasting , 1989, Conference Papers Power Industry Computer Application Conference.

[3]  Pu Wang,et al.  Electric load forecasting with recency effect: A big data approach , 2016 .

[4]  Ming-Wei Chang,et al.  Load Forecasting Using Support Vector Machines: A Study on EUNITE Competition 2001 , 2004, IEEE Transactions on Power Systems.

[5]  Michael P. Clements,et al.  A companion to economic forecasting , 2004 .

[6]  Carlos E. Pedreira,et al.  Neural networks for short-term load forecasting: a review and evaluation , 2001 .

[7]  M. Guha.  APA Dictionary of Psychology , 2007 .

[8]  Tao Hong,et al.  Short Term Electric Load Forecasting , 2012 .

[9]  Rob J Hyndman,et al.  Probabilistic energy forecasting: Global Energy Forecasting Competition 2014 and beyond , 2016 .

[10]  When One Size No Longer Fits All – Electric Load Forecasting with a Geographic Hierarchy Executive Summary , 2013 .

[11]  Tao Hong,et al.  A Naïve multiple linear regression benchmark for short term load forecasting , 2011, 2011 IEEE Power and Energy Society General Meeting.

[12]  Pierre Pinson,et al.  Global Energy Forecasting Competition 2012 , 2014 .

[13]  Pu Wang,et al.  Weather station selection for electric load forecasting , 2015 .

[14]  Rob J Hyndman,et al.  Short-Term Load Forecasting Based on a Semi-Parametric Additive Model , 2012, IEEE Transactions on Power Systems.