For the construction of multivariable regression models, stepwise covariate selection is a preferred strategy. Numerically stable and easy to apply, a model with fewer covariates. Different covariates are included in the model based on the different significance levels pre-specified by statisticians. Further research using these models could result in biases. This study presents a new method for selecting covariates for stepwise proportional hazards regression that does not require a significance level to be defined beforehand. For the final model selection, many models with varied numbers of covariates were generated. A SAS macro with a user-friendly interface was created. The final models can be determined by users of the macro based on estimated characteristic changes in the overall models, differences in the covariate effects on the response variable, and their unique requirements. Model selection is significantly easier with this strategy than with the intentional or best subsets methods. Stepwise covariate selection methods were enhanced using this strategy. A wide range of applications is envisioned.
Author(S) DetailsJian Sun
School of Public Health, University of Alberta, Edmonton, Canada and Department of Medicine, University of Calgary, Calgary, Canada.
View Book:- https://stm.bookpi.org/RDMMR-V9/article/view/4587
No comments:
Post a Comment