Time Series Forecasting for Decision Making on City-Wide Energy Demand: A Comparative Study
Orhan Nooruldeen, Salwan F. Alturki, Mohammed Rashad Baker, Ahmed Ghareeb
Abstract
Time series modeling and forecasting are critical in various practical applications, including the energy sector, and have been actively investigated in this field for several years. Many relevant methods for enhancing the accuracy and efficacy of time series modeling and forecasting have been proposed in the literature. This study aims to provide a comparative analysis of various common time series modeling and forecasting techniques for the daily electricity demand of the city of Kirkuk. The ability of the presented models to be extrapolated as well as increasing the confidence in models are also examined. Two years of out-of-sample data are used to validate the models. The Long Short-term Memory (LSTM) outperformed the other series types, demonstrating good agreement with the actual data. This study has implications for boosting renewable energy deployment, planning demand-side management, and measuring energy and cost-saving actions.