This paper proposes a least-square (LS) support vector machine (SVM)-based model for short-term solar power prediction (SPP). The input of the model includes historical data of atmospheric transmissivity in a novel two-dimensional (2D) form and other meteorological variables, including sky...