科研之友
微信
新浪微博
Facebook
分享链接
摘要
This paper presents an adaptive-network-based fizzy inference system (ANFIS) for long-term natural Electric consumption prediction. Six models are proposed to forecast annual Electric demand. 104 ANFIS have been constructed and tested in order to find the best ANFIS for Electric consumption. Two parameters have been considered in the construction and examination of plausible ANFIS models. The type of membership function and the number of linguistic variables are two mentioned parameters. Six different membership functions are considered in building ANFIS, as follows: the built-in membership function composed of the difference between two sigmoidal membership functions (dsig), the Gaussian combination membership gauss2), the Gaussian curve built-in membership gauss), the generalized bell-shaped built-in membership gbell), the H-shaped built-in membership pi), psig. Also, a number for linguistic variables has been considered between 2 and 20. The proposed models consist of input variables such as: Gross Domestic Product (GDP) and Population (POP). Six distinct models based on different inputs are defined. All of the trained ANFIS are then compared with respect to the mean absolute percentage error (MAPE). To meet the best performance of the intelligent based approaches, data are pre-processed (scaled) and finally our outputs are post-processed (returned to its original scale). The ANFIS model is capable of dealing with both complexity and uncertainty in the data set. To show the applicability and superiority of the ANFIS, the actual Electric consumption in industrialized nations including the Netherlands, Luxembourg, Ireland, and Italy from 1980 to 2007 are considered. With the aid of an autoregressive model, the GDP and population by 2015 is projected and then with yield value and best ANFIS model, Electric consumption by 2015 is predicted.
- 出版日期2010
