Support of randomForest and ranger
models which have been created using parsnip
Fix class checking for when formula notation is used in randomForest
Significant reduction in compute time for calculating false positive rates by sampling only unique selection frequencies
Addition of tidy tools (dplyr, tibble,
magrittr)
internals now implemented in C++ via
Rcpp thanks to Dr Jasen Finch (@jasenfinch)Implemented Strategy-1 from Konukoglu,E. and Ganz,M.,2014. Approximate false positive rate control in selection frequency for random forest
Support for randomForest and ranger
forest objects
Calculate selection frequency threshold for a given false positive rate (alpha)
False positive rate feature selection
Wrapper for selection frequencies extract from objects