Medium term electricity energy demand prediction in Turkey with adaptive neuro fuzzy inference system (ANFIS)

Due to the reasons of the specific structure of electricity demand, as electricity must be consumed as it is generated becuase it can not be stored, designing a system which can always meet the demand is the most important pillar of power system planning. In this context forecasting of electricity demand accurately and consistently has great importance. In the literature many studies on this subject have been made. Unlike other studies, in this thesis medium term gross electricity demand prediction for Turkey is made by using a combination of ANN (Artificial Neural Network) and fuzzy inference system which is called ANFIS. In order to measure the performance of the created model, a regression analysis model is used. Obtained estimation results are compared through the mean absolute error criteria. As a result of evaluations, it has been reached that ANFIS has provided a quite successful prediction performance.