An Extensive Simulation Study for Evaluation of Penalized Variable Selection Methods in Logistic Regression Model with High Dimensional Data

Variable selection, as a category of supervised methods, is a procedure in statistics that involves selecting a subset of important variables from a larger set of variables. The variable selection process in high dimensional data is quite significant to avoid overfitting and produces meaningful results from the model. Lasso, Elastic Net, Adaptive Lasso, and Adaptive Elastic Net, known as penalized methods, are frequently used methods for variable selection to reduce dimensionality in the logistic regression model with high dimensional data. This research aims to examine and compare the performances of these penalized methods in the variable selection process in logistic regression under different scenarios through an extensive simulation study in high- dimensional data.