This chapter discusses the application of regression techniques to a hypothetical problem of infant mortality, to zero in factors that might be responsible for the outcome of the problem. LASSO is good for selecting independent variables that are responsible for an output, but it fails to capture a set of correlated variables. This issue is overcome by the elastic-net formulation. It is a good option when the variables are known to be correlated but the exact group of correlated variables is unknown. The chapter also discusses multi-task regression to explain a variety of outputs for the hypothetical problem.