Stochastics and Statistics Seminar

Slope meets Lasso in sparse linear regression

Speaker Name: Pierre Bellec (Rutgers)

Date: February 10, 2017

Time: 11:00am

Location: E18-304

Abstract:

We will present results in sparse linear regression on two convex regularized
estimators, the Lasso and the recently introduced Slope estimator, in the
high-dimensional setting where the number of covariates p is larger than the
number of observations n. The estimation and prediction performance of these
estimators will be presented, as well as a comparative study of the assumptions
on the design matrix.

https://arxiv.org/pdf/1605.08651.pdf
[Joint work with Guillaume Lecue and Alexandre B. Tsybakov]

Speaker Bio:

I am an Assistant Professor of statistics at Rugers, the state university of New Jersey. I obtained my PhD in 2016 from ENSAE ParisTech, where I was fortunate to have Alexandre Tsybakov as my PhD advisor. My research interests include aggregation of estimators, shape restricted regression, confidence sets, high-dimensional statistics and concentration inequalities.