부산대학교 도서관

Solar energy generated by the sunlight is non schedulable due to the stochastic behavior of meteorological conditions. Also the equipment for measuring the solar irradiance is expensive and rarely available in all locations of the globe. Hence prior knowledge of solar radiation is very important, for better management, sizing and control of solar energy installations. Predictive data mining such as Artificial Neural Network (ANN) is one of the most reliable and accurate model for forecasting solar radiation. There are a number of meteorological and geographical parameters which affect the solar radiation prediction model. So, identification of most influential parameters for better prediction accuracy is a crucial factor. In this paper, three attribute evaluators of Waikato Environment for Knowledge Analysis (WEKA) such as Classifier Subset Eval, CFS Subset Eval and Wrapper Subset Eval are used to select the most influential input parameters for ANN model for prediction of solar radiation in different meteorological locations of Bangladesh. Then the selected parameters are used to train, validate and test the ANN model. During the training process the number of hidden layer neuron are changed to find optimal network architecture. The prediction accuracy of the ANN models are evaluated using the statistical parameters such as Mean Absolute Percentage Error(MAPE), Mean Square Error(MSE) and goodness of fit(R). The simulated result shows that the subset selected by the Wrapper subset evaluator outperforms.