Bootstrapping the coefficients of multiple logistic regression model in medicine data

Abstract

Bootstrap is becoming a useful and very popular tool for obtaining estimations and confidence intervals for coefficients in many researches in different scientific fields without making assumptions about the population. Our goal is to apply bootstrap technique in parameter estimation and confidence intervals for the coefficients in Multiple Logistic Regression model in a study using medical records. We will use R programming language and SPSS software to obtain the coefficients of the model and the estimations using non-parametric bootstrap and we will also make a comparison of the results emphasizing the importance of using resampling methods even in a study with real data.