Grouped Gaussian Processes for Solar Power Prediction

Astrid Dahl, Edwin V. Bonilla

arXiv preprint · 07 June 2018

In one sentence. Coupled Gaussian process priors improve multi-site solar power and wind speed forecasts, matching or beating benchmarks with better uncertainty quantification.

Abstract

We consider multi-task regression models where the observations are assumed to be a linear combination of several latent node functions and weight functions, which are both drawn from Gaussian process priors. Driven by the problem of developing scalable methods for forecasting distributed solar and other renewable power generation, we propose coupled priors over groups of (node or weight) processes to exploit spatial dependence between functions. We estimate forecast models for solar power at multiple distributed sites and ground wind speed at multiple proximate weather stations. Our results show that our approach maintains or improves point-prediction accuracy relative to competing solar benchmarks and improves over wind forecast benchmark models on all measures. Our approach consistently dominates the equivalent model without coupled priors, achieving faster gains in forecast accuracy. At the same time our approach provides better quantification of predictive uncertainties.

Labels

Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.