A Robust Variable Selection Method for Sparse Online Regression via the Elastic Net Penalty

Variable selection has been a hot topic, with various popular methods including lasso, SCAD, and elastic net. These penalized regression algorithms remain sensitive to noisy data. Furthermore, “concept drift” fundamentally distinguishes streaming data learning from batch learning. This article presents method for noise-resistant regularization variable in streams multicollinearity, dubbed canal-adaptive net, which is similar net encourages grouping effects. In comparison the canal adaptive especially advantageous when number of predictions (p) significantly larger than observations (n), are multi-collinear. Numerous simulation experiments have confirmed that higher prediction accuracy ridge regression, multicollinearity noise.