GSO Spring Speaker 2011
The Lasso: Some Novel Algorithms and Applications
Rob Tibshirani
Professor
Department of Statistics, Stanford University
Venue: LWSN 1142
Abstract:
I will discuss some procedures for modeling high-dimensional data, based on L1 (lasso)-style penalties. I will describe pathwise coordinate descent algorithms for the lasso, which are remarkably fast and facilitate application of the methods to very large datasets for the first time. I will then give examples of new applications of L1 penalties to microarray classification, the fused lasso for signal detection and the matrix completion problem.