Medium-Term Load Forecasting Model Using Radial Basis Function Algorithm

This study developed two medium-term electricity load forecasting models for Banda Aceh city using Radial Basis Function method and Artificial Neural Network. There are three load groups used to create the model, including social, household and industrial load groups. The dataset used consists of 4 (four) input variables and one target. This research developed 3 training functions that were applied to both models to obtain the best forecast model. The training functions include Lavenberg-Marquardt, Gradient Descent and Adaptive Learning Rate, and Gradient Descent with momentum and adaptive learning rate. Based on the simulation and its MAPE value, the radial basis function algorithm is highly suitable for implementing mid-term electricity load forecasting modeling.