Conversion of quarterly input data to demand prediction with high accuracy using adaptive neuro-fuzzy inference system: The case of Turkey

Energy is the most basic input in the production process for the realization of social and economic development. Unstockable nature of electricity energy necessitates designing a system that can always meet the demand. This is the most important step of energy system planning. Medium term prediction is important especially for Energy Systems Management in allocating production capacity, market research and network maintenance planning. Adaptive Neuro Fuzzy Inference System (ANFIS) is used for this study with the compilation of the obtained three-month data for twenty years. ANFIS has obtained results with high accuracy versus regression analysis even for crisis periods because of its adaptive architecture covering the whole dynamic system structure. Gross electricity demand could be predicted with % 96.74 accuracy by ANFIS as compared previous studies.

کلیدواژه ها:

Adaptive neuro-fuzzy systems ، electricity demand forecasting ، medium-term ، seasonal fluctuations