nielsaka/modelconf: Estimation of Model Confidence Sets

This package offers an implementation of the algorithms developed by Hansen, Lunde and Nason (2011). The authors introduce the notion of a model confidence set (MCS) to account for the degree of uncertainty surrounding model selection. In analogy to confidence intervals for population parameters, the MCS methodology estimates a set of models that is expected to contain the best model(s) with a given probability.

install.packages("remotes")
remotes::install_github("nielsaka/modelconf")