A Computational Statistics: Multilayer Feed-Forward Neural Network Approach To Two-Way Anova Toward Linear Model

Abstract

This study investigates the connection between Two-Factor Analyses of Variance (Two-way ANOVA) and multiple linear regression (MLR). This paper introduces straightforward instructions and a template for data transformation, making it an essential and mandatory step in constructing multiple linear regression models. The primary concept involves converting Two-way ANOVA into an applied linear model, providing a valuable framework for establishing a comprehensive connection between Two-way ANOVA and linear models. The step-by-step process required to fit the linear model is illustrated in this paper, along with a validation procedure using a Multilayer Feed-Forward Neural Network (MLFFNN). The application of a response surface plot aimed to elucidate the interplay between smoking and gender factors concerning uric acid characteristics. Two-way ANOVA and multiple linear regression (MLR) are interrelated; therefore, the multiple linear regression method is an alternate data analysis strategy.