Seminar by Abbas Khalili

Sparse estimation in heterogeneous varying coefficient regression models

In this talk, we will present statistical methodologies based on regularized local likelihood for sparse estimation in mixture of varying coefficient regressions. These models are useful for estimating the effects of a set of covariates on a response variable where there are unknown underlying subpopulations and also the effects may vary according to an index variable such as time or location. Although complex, this situation frequently occurs in real data applications which we demonstrate using a genetic dataset. We will discuss statistical properties of the methods, and we also evaluate their finite-sample performance via simulations. We will apply the proposed methods to a genetic dataset. If time permits, we discuss some future directions!