sparsefreg: sparsefreg: A package for performing scalar-on-function...

Description

Description

The sparsefreg package implements the MISFIT (Multiple Imputation for Sparsely-sampled Functions at Irregular Times) approach, used for scalar-on-function regression when the functional covariate is observed sparsely. MISFIT takes a missing data approach to imputing (over a denser grid) a functional covariate which is only observed sparsely on a grid. More specifically, the curves are represented by their Karhunen-Loeve expansion where the scores are imputed and can be used to reconstruct the curves, and/or directly used to fit an FPC regression model.


justin-petrovich/sparsefreg documentation built on Aug. 20, 2020, 9:04 p.m.