The Lasso: 22 years later: (Tuesday; Berkeley-Stanford joint colloquium)

Seminar | April 11 | 4-5 p.m. | 277 Cory Hall

 Rob Tibshirani, Stanford University

 Department of Statistics

In April 1995 I gave the Stanford-Berkeley seminar entitled "Regression Shrinkage and Selection via the Lasso". I will recount that day and review what has happened in this area of research since that time. I will also discuss some new developments (by others) in the computation of best subsets regression, a main competitor to the lasso, and present the results of a large scale numerical study that sheds new light on the relative merits of these and other approaches. Finally, I will update my (somehat well-known?) slide comparing Berkeley and Stanford Statistics that was presented at that 1996 seminar, without, hopefully, offending too many friends in both departments.

The numerical study presented is joint work with Trevor Hastie and Ryan Tibshirani