In the context of finding the "best" predictive model, this research project focuses on assessing and fitting models with least-absolute-deviation techniques.