A New Approach for Parameter Estimation in Fuzzy Logistic Regression

Abstract

Logistic regression analysis is used to model categorical dependent variable. It is usually used in social sciences and clinical research. Human thoughts and disease diagnosis in clinical research contain vagueness. This situation leads researchers to combine fuzzy set and statistical theories. Fuzzy logistic regression analysis is one of the outcomes of this combination and it is used in situations where the classical logistic regression assumptions' are not satisfied. Also it can be used if the observations or their relations are vague. In this study, a model called Fuzzy Logistic Regression Based on Revised Tanaka's Fuzzy Linear Regression Model is proposed. In this regard, the methodology and formulation of the proposed model is explained in detail and the revised Tanaka's regression model is used to estimate the parameters. The Revised Tanaka's Regression model is an extension of Tanaka's Regression Model in which the objection function is developed. An application is performed on birth weight data set. Also, an application of diabetes data set used in Pourahmad et al.'s study was conducted via our proposed data set. The validity of the model is shown by the help of goodness of fit criteria called Mean Degree Memberships (MDM).