Paper Info
Title
A Forward Variable Selection Method for Fuzzy Logistic Regression
Abstract
The logistic regression analysis is a popular method for describing the relation between variables. However, when there are a big number of variables in the regression model, the selection of the best model becomes a major problem. In this condition, the question is which subset of predictors can best predict the response pattern, and which process can be used to achieve such a subset. This article is written to answer this questioning fuzzy logistic regression models. To this end, based on the existing criteria of regression models, three goodness-of-fit criteria, namely MSEF, AICF, and $$C_{p}^{\text{F}}$$ , are proposed. These criteria are helpful to select the best-fitted model among all possible fuzzy logistic regression models with fuzzy covariates and responses. In addition, based on the concepts of efficiency level and MSEF, a forward model selection method for fuzzy logistic regression is proposed. The proposed method is justified by some simulation studies, indicating the good performance and efficiency of the method. In addition, we applied the presented methods in a clinical trial study.
Year
DOI
Venue
2019
10.1007/s40815-019-00615-z
International Journal of Fuzzy Systems
Keywords
Field
DocType
Fuzzy logistic regression, Fuzzy predictor, Forward method, Goodness-of-fit criteria
Covariate,Mathematical optimization,Feature selection,Regression analysis,Fuzzy logic,Model selection,Statistics,Logistic regression,Mathematics
Journal
Volume
Issue
ISSN
21
4
1562-2479
Citations 
PageRank 
References 
0
0.34
0
Authors
3
Name
Order
Citations
PageRank
Fatemeh Salmani100.34
Seyed Mahmoud Taheri281.64
Alireza Abadi361.83