Ertugrul, EdipSahin, MehmetAggun, Fikri2024-12-242024-12-242015978-1-4673-7386-92165-0608https://hdl.handle.net/20.500.12604/576123nd Signal Processing and Communications Applications Conference (SIU) -- MAY 16-19, 2015 -- Inonu Univ, Malatya, TURKEYSolar energy, which is clean and renewable energy source, is a popular subject. The estimation of solar radiation can be done instead of long term measurements. Therefore, the satellite and meteorological values of 53 different locations of Turkey were used for estimations of solar radiation. In this study a hybrid approach was proposed. The train dataset was reduced by employing two times similarity and the reduced dataset was utilized with support vector machine to predict global solar radiation. Additionally, the proposed method was validated by employing neural network, linear regression, k nearest neighbor, extreme learning machine, Gaussian process regression and kernel smooth regression. This study was showed that the machine learning methods can be used instead of long term measurement before investments.trinfo:eu-repo/semantics/closedAccessSolar RadiationMachine LearningRegressionConference Object16111614N/AWOS:000380500900383