Why `mgcv` is awesome


The mgcv package introduces a flexible approach to regression models by fitting various penalized smoothers. This includes splines and related methods that may be a more appropriate approximation for non-linear relationships or noisy data. This talk introduces the subject matter in general, visualizes how these models might fit sigmoidal relationships, introduces the smoothing penalty and how to estimate it, as well as showing how to fit a Generalized Additive Model using mgcv in R.

Jan 21, 2020 5:30 PM — 7:30 PM