Browse Teach Curate

Penalized Weighted GEEs for High-Dimensional Longitudinal Data with Informative Cluster Size

A regression method for longitudinal data with many covariates that fixes the bias generalized estimating equations suffer when cluster size is informative (related to the outcome): clusters are reweighted to remove that bias and a penalty is added for variable selection in high dimensions, with proofs that the estimator is consistent in both model selection and estimation and asymptotically equivalent to the oracle estimator.

Demerzel Text

Penalized weighted GEEs for high-dimensional longitudinal data with informative cluter size Yue

Longitudinal data consist of repeated, correlated measurements on the same subjects or clusters over time, and generalized estimating equations (GEEs) are a standard way to fit regression models to t…