Comparison between some methods for estimation and variables selection for semi parametric additive model with practical application

Abstract EN

The additive partial linear model (APLM) was used to estimate the effects of some linear and non-linear explanatory variables on the response variable represented in the market value of the Baghdad soft drinks company.

four methods have been to choose and estimate the model, namely ALasso, SCAD, elastic net and adaptive elastic net.

the methods used were compared using comparison criteria represented by mean squares error and coefficient of determination.

through the results of the analysis, it has been noticed that the method adaptive elastic net is more efficient than to other methods.