Özet:

Overestimation of energy usage would result wasted capacity and financial loss while underestimation would result devastating power disruptions. The objective is therefore to simultaneously improve the efficiency of energy systems and develop state-of-art techniques which accurately predict future energy demand. This is especially vital for developing regions such as Northern Cyprus. In this study, we provide medium and long term electricity consumption models using varying time-series and associative forecasting techniques. When time series methods are studied, a novel approach of optimizing initial parameters in addition to smoothing parameters is applied. A benchmark study is provided to evaluate the performance of different consumption models in terms of forecast accuracy

Anahtar Kelimeler: