Parameter Estimation Based Type-II Fuzzy Logic and Comparison with Robust Methods

Authors : Kamile ŞANLI KULA, T E DALKILIÇ

Pages : 118-124

Doi:10.36753/mathenot.421420

View : 31 | Download : 12

Publication Date : 2016-04-15

Article Type : Research Paper

Abstract :Parameter estimation is one of the important stages of regression analysis. In the regression analysis, while parameter estimation by classical methods there are a number of assumptions need to be satisfied. One of them is error are normally distributed. In the case that the data sets have outliers, providing of this assumption becomes more difficult. When a data set has outliers, robust methods such as the M method insert ignore into journalissuearticles values(Huber, Hampel, Andrews and Tukey); are used for estimating parameters. In this paper the Adaptive Network Based Fuzzy Inference System insert ignore into journalissuearticles values(ANFIS); is used to parameter estimation which is the neural network architecture based type-II fuzzy logic. The proposed method has the properties of a robust method, because the process does not give permission to the intuitional and is not affected by the outliers. Consequently, another aim of this study is, to compare the proposed method with the robust methods that are mentioned above. 
Keywords : Type II fuzzy logic, outliers, robust regression, membership function

ORIGINAL ARTICLE URL