Maximum penalized likelihood

This dissertation is composed of three papers which address the problem of variable selection for models with missing data. In the first paper, we consider variable selection for generalized linear models with missing data, including missing covariate and/or response data. The s…