Open AccessLoad forecasting by regression model based on fuzzy rules
Author(s) -
Binh Phan,
Manh Luong
Publication year - 2014
Publication title -
khoa học công nghệ
Language(s) - English
Resource type - Journals
ISSN - 1859-0128
DOI - 10.32508/stdj.v17i1.1267
Subject(s) - logarithm , linear regression , mathematics , correlation coefficient , cluster analysis , proper linear model , fuzzy logic , function (biology) , linear model , population , econometrics , statistics , mathematical optimization , computer science , artificial intelligence , polynomial regression , mathematical analysis , demography , evolutionary biology , sociology , biology
The forecasting models by traditional regression function have the crisp functions such as Y=f(x1, x2 ,….,xn) or logY=f(logx1, logx2 ,….,logxn). Here f has the linear form and xi are the factors such as GDP, temperature, industrial output, population… But these models are able to be used only when the linear correlation existed (expressed by the correlation coefficient). This paper introduced the regression model based on the fuzzy Takagi-Sugeno rules. These rules are built by using the subtractive clustering. The model is used for the general case, even when there are no the crisp function f. Examining shows that the good results are obtained in the case of traditional correlation such as linear or linear by logarithm. The results are also satisfactory for the case of unknown correlation. The electricity consumption forecasting due to the temperature factor for one substation of HochiMinh city was carried out.