Speaker: Bereket Kindo, PhD, Humana

Title: "Bayesian estimation of sums of sparse and dense signals"

In the linear regression model setting, and regularization methods are used for the recovery of sparse and dense signals, respectively. In this talk, we will outline a Bayesian approach for the estimation of sums of sparse and dense signals. That is, we assume the estimand consists of a sparse component with few strong signals, and a dense component with many small signals. We also explore the utility of the sparse + dense estimator using simulated examples and real data application.

Stay connected TwitterFacebookLinkedInYouTubeInstagram