Applied Math Talk: Variable Selection via Arbitrary Rectangle-Range Generalized Elastic Net given by Yujia Ding (CGU)

We propose a regularization and variable selection method, named arbitrary rectangle-range generalized elastic net (ARGEN). It can be applied in high dimensional sparse linear regression models. We propose an algorithm to solve ARGEN; it is an extension of multiplicative updates. Multiple simulation studies and a real-world application in the stock market show that ARGEN applies to more complicated problems, outperforms and extends the lasso, ridge, and elastic net.