Random Forest Ensemble of Support Vector Regression Models for Solar Power Forecasting

To mitigate the uncertainty of variable renewable resources, two off-the-shelf machine learning tools are deployed to forecast the solar power output of a solar photovoltaic system. The support vector machines generate the forecasts and the random forest acts as an ensemble learning method to combine the forecasts. The common ensemble technique in wind and solar power forecasting is the blending of meteorological data from several sources. In this study though, the present and the past solar power forecasts from several models, as well as the associated meteorological data, are incorporated into the random forest to combine and improve the accuracy of the day-ahead solar power forecasts. The performance of the combined model is evaluated over the entire year and compared with other combining techniques.

💡 Research Summary

The paper proposes a novel ensemble framework for day‑ahead solar photovoltaic (PV) power forecasting that combines multiple Support Vector Regression (SVR) models with a Random Forest (RF) meta‑learner. Recognizing that most existing solar forecasting studies either blend meteorological inputs from several sources or apply a single machine‑learning model, the authors introduce a two‑stage approach that also leverages the forecasts generated by the individual SVR models as additional features for the ensemble.

Data were collected from a 1.56 kW PV system located in southern Australia, covering the period from April 2012 to May 2014. The dataset includes hourly measured PV output and corresponding Numerical Weather Prediction (NWP) variables. To create diversity among the base learners, the authors split the historical data into two subsets: the full 26‑month record and the most recent 12‑month window. For each subset they train 12 SVR models, varying (i) the data normalization technique (either min‑max scaling to