Point and interval forecasting of solar irradiance with an active Gaussian process

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

5 Scopus Citations
View graph of relations


Related Research Unit(s)


Original languageEnglish
Pages (from-to)1020-1030
Journal / PublicationIET Renewable Power Generation
Issue number6
Online published3 Mar 2020
Publication statusPublished - 27 Apr 2020


A Gaussian process regression (GPR) with active learning is proposed for developing the solar irradiance point andinterval forecasting models, which consider the spatial-temporal information collected from a targeted site and a number ofneighbouring sites. To enhance the performance of the GPR-based model an active learning process is developed forconstructing an ad-hoc input feature set, selecting training data points, and optimising hyper-parameters of GPR models. Tovalidate the advantages of the proposed method, a comprehensive computational study is conducted based on solar irradiancedata collected from the northwest California area. In the point forecasting, the proposed method beats the state-of-the-artbenchmarking methods including classical statistical models and data-driven models according to values of the normalised rootmean squared error, normalised mean absolute error, normalised mean bias error, and coefficient of determination. In theinterval forecasting, the proposed method outperforms the persistence model, autoregressive model with exogenous inputs,generic GPR, as well as two recently reported forecasting methods, the bootstrap-based extreme learning machine and quantileregression, in terms of the forecasting reliability. Computational results show that the proposed method is more effective thanwell-known existing benchmarks in the point and interval forecasting of the solar irradiance.

Research Area(s)

  • mean square error methods, regression analysis, Gaussian processes, learning (artificial intelligence), autoregressive processes, solar power, feedforward neural nets, power generation planning, power engineering computing, interval forecasting, spatial-temporal information, solar irradiance forecasting, active learning process, ad-hoc input feature set, solar irradiance data, point forecasting, statistical models, data-driven models, normalised mean absolute error, normalised mean bias error, persistence model, autoregressive model, forecasting methods, bootstrap-based extreme learning machine, forecasting reliability, active Gaussian process regression, northwest California, normalised root mean squared error, coefficient of determination, quantile regression, solar power generation, JAYA ALGORITHM, PREDICTION, GENERATION, MODEL, WIND, NETWORK