Stability of Variable Selection Methods: A Bootstrapped Comparison of Ssvs, Lasso, and Stepwise Regression

We compared predictive stabilities of Stochastic Search Variable Selection (SSVS), Least Absolute Shrinkage and Selection Operator (LASSO), and stepwise regression using a bootstrapped dataset of eight behavioral factors predicting affect and stress and found that SSVS is the most stable method, followed by LASSO and stepwise regression.