Comparative Analysis of Neural Network and Radial Basis Functions Approaches for Electricity Consumption Forecasting in Five Different Regions in Albania

Forecasting electricity consumption remains essential, necessitating the application of recent, hybrid, and classical models. Traditional approaches often fall short in capturing nonlinear and irregular data patterns. This paper compares three distinct approaches for surrogate modeling and forecasting daily electricity consumption across five regions in Albania: a Radial Basis Function approach, a feedforward Neural Network approach, and a Recurrent Neural Network approach. Our aim is to develop a generalized model for these regions and provide highly accurate one-day-ahead predictions, considering daily average temperature as an exogenous variable. We conduct a comparative analysis based on model fit, computational efficiency, and error measurement metrics. Our findings underscore the exceptional performance of the RBF ARX approach, showcasing its efficacy in accurately forecasting electricity consumption based on daily data. Daily mean temperature is identified as a key factor influencing model performance.