Multiple Regression: Methodology and Applications

Abstract

Multiple regression is one of the most significant forms of regression and has a wide range of applications. The study of the implementation of multiple regression analysis in different settings contributes to the development of relevant theories and the improvement of models. In this paper, four different kinds of regressions are discussed individually by referring to different articles. The four kinds of regressions discussed are multivariable/multiple linear regression, multivariate multiple linear regression, multinomial logistic regression, and multivariate non-linear regression. As for multivariable/multiple linear regression, examples in the manufacturing industry and medical field show that it can be applied in more fields. Multivariate multiple linear regression is more accurate than multivariable/multiple linear regression and can be used with more than a variable. Multinomial logistic regression is relatively mature and accurate, and can help people well solve the problem with non-linearity and multiple independent variables. It does not require the variables to obey a multivariate normal distribution, and is more widely used as well. Multivariate non-linear regression, however, cannot be used properly without powerful professional knowledge. This paper investigates the theoretical development and model applications of multiple regression to demonstrate the flexibility and broadness of the adoption of multiple regression analysis.