An Integrated approach for fuzzy logistic regression

The aim of this study is to introduced an integrated fuzzy logistic regression approach to describe the relationship between crisp inputs and fuzzy binary output. For this reason, we integrated the fuzzy logistic regression methods proposed by Pourahmad et al. [17] and Sohn et al. [24] to define a possibility measure for each case and then used the logarithmic transformation of possibilistic odds as fuzzy output observations. To estimate the parameters of the fuzzy logistic regression model, Diamond [5]’s Fuzzy Least Squares (FLS) approach is used. A numerical example is presented and obtained results are compared with classic logistic regression model