A New Fuzzy Regression Model by Mixing Fuzzy and Crisp Inputs
Magda M. M. Haggag

This paper proposes a new form of the multiple regression model (mixed model) based on adding both fuzzy and crisp input data. The least squares approach of the proposed multiple regression parameters are derived in different cases. This derivation is based on the fact that each fuzzy datum is a nonempty compact interval of the real line. The main contribution is to mix both fuzzy and crisp predictors in the linear regression model. The mixed fuzzy crisp model will be introduced mathematically and by coded via R-language. The least squares of the regression parameters will be derived and evaluated using distance measures. Numerical examples using generated data showed best results for the mixed fuzzy crisp multiple regression models compared to the multiple fuzzy models.

Full Text: PDF     DOI: 10.15640/arms.v6n2a2