TY - JOUR

T1 - Intuitionistic fuzzy ridge regression functions

AU - Kizilaslan, Busenur

AU - Egrioglu, Erol

AU - Evren, Atif Ahmet

PY - 2020/3/1

Y1 - 2020/3/1

N2 - Developing technology shows how useful fuzzy inference systems in lots of applications. Fuzzy functions approach which is one of the important fuzzy inference system for time series forecasting. In fuzzy functions approach, the membership values and their non-linear transformations are used together with original input variables to increase the prediction power. However, multicollinearity problem can be occured because of using these correlated variables. Purpose of the paper is to propose a new fuzzy forecasting method with intuitionistic fuzzy sets which has addition information known as hesitation degree. In this case, both intuitionistic fuzzy sets and their non-linear transformations is used to increase the prediction power. Ridge regression method is preferred to obtain intuitionistic fuzzy functions without exposed to multicolinearity problem. To demonstrate the performances of proposed method, some real world time series data are used and the results have shown that the effectiveness of the proposed method in conrast to other methods.

AB - Developing technology shows how useful fuzzy inference systems in lots of applications. Fuzzy functions approach which is one of the important fuzzy inference system for time series forecasting. In fuzzy functions approach, the membership values and their non-linear transformations are used together with original input variables to increase the prediction power. However, multicollinearity problem can be occured because of using these correlated variables. Purpose of the paper is to propose a new fuzzy forecasting method with intuitionistic fuzzy sets which has addition information known as hesitation degree. In this case, both intuitionistic fuzzy sets and their non-linear transformations is used to increase the prediction power. Ridge regression method is preferred to obtain intuitionistic fuzzy functions without exposed to multicolinearity problem. To demonstrate the performances of proposed method, some real world time series data are used and the results have shown that the effectiveness of the proposed method in conrast to other methods.

KW - Fuzzy Functions

KW - Intuitionistic Fuzzy Sets

KW - Multicollinearity

KW - Ridge Regression

KW - Time Series Forecasting

U2 - 10.1080/03610918.2019.1626887

DO - 10.1080/03610918.2019.1626887

M3 - Journal article

AN - SCOPUS:85067491928

VL - 49

SP - 699

EP - 708

JO - Communications in Statistics: Simulation and Computation

JF - Communications in Statistics: Simulation and Computation

SN - 0361-0918

IS - 3

ER -