Robert Tibshirani

Overview

Robert Tibshirani is a statistician at Stanford University (formerly University of Toronto) known for foundational contributions to statistical learning — the lasso, generalized additive models (with Hastie), and the bootstrap/statistical-learning textbooks.

Key Contributions / Features

• The lasso — Tibshirani (1996): Introduced $\ell_1$-penalized regression as a method that shrinks and selects simultaneously, producing sparse interpretable models with ridge-like stability. See Lasso and Tibshirani (1996).
• Generalized additive models (Hastie–Tibshirani 1990): Flexible additive nonparametric regression via smoothers and backfitting.
• glmnet / coordinate descent (Friedman-Hastie-Tibshirani 2010): fast coordinate-descent path algorithms for penalized GLMs.
• Textbooks: The Elements of Statistical Learning (with Hastie and Friedman) and An Introduction to Statistical Learning — standard references for modern statistical/machine learning.

Related

• Lasso
• Variable Selection
• Coordinate Descent
• Tibshirani (1996)
• Friedman-Hastie-Tibshirani (2010) Regularization Paths for Generalized Linear Models via Coordinate Descent
• Jerome Friedman
• Trevor Hastie