engleski

Appropriateness and Limitations of Multiple Regression Analysis Usage in Biomedical Research

INTRODUCTION: Best model selection is a usual problem when researcher has to deal with relatively large number of observed variables. Especially, in multiple regression analysis, where goal is identification of relations between single variable (criterion) and multiple variables (predictors) there exists different approaches to model generation. MATERIALS AND METHODS: Appropriateness and limitations of usage of multiple regression analysis is discussed from aspects of practice in biomedical sciences. CONCLUSION: Recommendations for adequate usage of multiple regression analysis in biomedical research are given.

multiple regression, optimal selection, stepwise algorithms, cross-validation

nije evidentirano