The prime reason for not using penalized linear regression is that you might get
better performance with another technique, such as an ensemble method. ensembles perform best in complicated problems (for example,
highly irregular decision surfaces) with plenty of data to resolve the problem’s complexities. In addition, ensemble methods for measuring variable importance can yield more information about the relationship between attributes and predictions. For example, ensembles will give second-order (and higher) information about what pairs of variables are more important together than the sum of their individual importance. That information can actually help squeeze more performance out of penalized regression.
When to Use Ensemble Methods page124
原文:http://www.cnblogs.com/hyqxln/p/6507444.html