چكيده به لاتين
The study of energy demand has always been of interest from different points of view, including macroeconomic, planning and policy making ones.
Electricity power is among the most important source of energy and is an indispensable factor in the development of a nation’s industries. The study of the economic aspects of the power industry is also of particular importance in Iran, as it can play an important role in the development process of the country in the future. In order to plan for the energy sector, including electricity, it is necessary to recognize consumer behavior. Demand function is one of the tools that helps us identify customer behavior by analyzing the effective factors and their impact on energy demand. On the other hand, electricity consumption planning has always been at center of considerations due to its high application in different sectors such as household and industry. So far, several methods have been proposed for modeling and predicting the electricity consumption of different subsections, including econometric methods and neurological and meta-innovative algorithms.The method used in this thesis is to introduce uncertainty in the modeling of the electric energy demand function, which is realized by using fuzzy linear regression with symmetric triangular fuzzy parameters.
In this study, fuzzy linear regression performance has been compared with a classical regression approach. Also, three indicators of predicted error mean square, mean prediction error and the mean absolute magnitude of the predicted error have been applied to compare fuzzy regression with classical regression, resulting in a more reliable and accurate prediction by fuzzy regression rather than classical regression. Domestic consumption was chosen as the case study, covering a time span of 1979 to1974. Furthermore, Gross domestic product, population and dom with domestic consumption as the dependent variable
