Model-Independent Variable Selection via the Rule-Based Variable Priority
Cross-Validated Cutoff Value for Variable Priority (VarPro)
Calculate VarPro Importance
Identify Anomalous Data
Individual Variable Priority (iVarPro): Case-Specific Variable Importa...
Model and subsapce aware out-of-distribution (OOD) scoring with outPro
Partial Effects for Variable(s)
Partial Plots for VarPro
Prediction for Isopro for Identifying Anomalous Data
Prediction on Test Data using Unsupervised VarPro
Prediction on Test Data using VarPro
Internal Utility Functions
Unsupervised Variable Selection using Variable Priority (UVarPro)
Show the NEWS file
Model-Independent Variable Selection via the Rule-based Variable Prior...
Obtain Strength Array and Other Values from a VarPro Object
A new framework of variable selection, which instead of generating artificial covariates such as permutation importance and knockoffs, creates release rules to examine the affect on the response for each covariate where the conditional distribution of the response variable can be arbitrary and unknown.