Global solar radiation forecast using an ensemble learning approach
In one sentence. Random forest and XGBoost ensembles predict site-specific solar radiation from station weather forecast reports, then convert it to PV power output.
Abstract
With the increase in demand for solar power, a solar power forecasting model is of maximum importance to allow a higher level of integration of non-conventional energy into the existing electricity grid. With the advancement in data availability, there’s a good time to use data-driven algorithms for enhanced prediction of solar energy generation. Gathering and analyzing data can predict solar energy generation and mitigate the impact of solar intermittency. During this research, we explore automatically creating prediction models that are site-specific utilizing machine learning to generate solar radiation from meteorological station weather forecast reports, and from the predicted solar radiation corresponding solar power output can be calculated depending upon the characteristics of the solar PV system used. The challenge is to enhance the accuracy of the forecast. Ensemble techniques like random forest (RF) and extreme gradient boosting (XGBoost) are well suited for solar radiation prediction as they improve stability as well as combine several machine learning models to reduce variation and bias which outperforms the majority of models, as a result making them a perfect model in the field of solar energy prediction.
Labels
- Topic
- Applications & impacts, Local & station forecasting
- Type
- Application study
- Method
- Classical ML (trees, GP, linear, kernels)
- Time scale
- 6 h–3 days
- Domain
- Local / station
- Data
- Station / in-situ, NWP analysis or forecasts (IFS, GFS, ...)
- Status
- Research
- Identifiers
- DOI:10.11591/ijpeds.v14.i1.pp496-505 OpenAlex:W4313654558
- Added
- 2026-10-22
Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.