nadir (noun): nā-dir
the lowest point.
{nadir} implements the super learner algorithm1. To quote the
Guide to SuperLearner2 (a previous implementation):
SuperLearner is an algorithm that uses cross-validation to estimate the performance of multiple machine learning models, or the same model with different settings. It then creates an optimal weighted average of those models, aka an “ensemble”, using the test data performance. This approach has been proven to be asymptotically as accurate as the best possible prediction algorithm that is tested.
Fitting with the minimum loss based estimation
literature,3,4 {nadir} is an implementation of the
super learner algorithm with improved support for flexible formula based
syntax and which is fond of functional programming techniques such as
closures, currying, and function factories.
Installation Instructions
{nadir} is available on CRAN – You can install it from CRAN using:
install.packages("nadir")The development version of {nadir} may be installed from GitHub via
running:
devtools::install_github("ctesta01/nadir")
Why {nadir} and why reimplement super learner again?
In previous implementations
({SuperLearner},
{sl3},
{mlr3superlearner}),
support for flexible formula-based syntax has been limited, instead
opting for specifying learners as models on an