Load forecasting via manifold regularization

The ability to accurately forecast the load plays an important role in electric power system planning and operating. In this paper, a novel approach was proposed for the electricity load forecasting by applying the manifold regularization learning methodology. Unlike traditional methods for load forecasting, the prediction method based on manifold regularization allows us to effectively exploit the geometric manifold structure of electricity load data in a semi-supervised learning setting. The effectiveness of the proposed approach is illustrated through an application to actual load data from the northwest China region.