Selection of network architecture and input sensitivity analysis for a Neural Network Energy Prediction Model

The focus of this article is to select the best architecture for a Neural Network Energy Prediction Model (NNEPM). A few network architecture is simulated and modeled; Multilayer Perceptron (MLP), Radial Basis Function (RBF), Generalized Radial Basis Function (GRBF), and Elman Network (Elman). From these networks, the network performances are compared and the best architecture is chosen for the NNEPM. The sensitivity of the inputs is also analyzed to verify the correlation and relationship of inputs and output of the NNEPM. In this study, NNEPM is analyzed on the sensitiveness of the model using different sets of data input to the model. Data inputs are categorized in several sets of condition with one of the inputs is given ±10% variations. All combinations of inputs are investigated and the sensitivity of the model is verified. The selected network architecture is the MLP to be simulated to give the best result and performance; root mean square error (RMSE).