nruPredict: miscellaneous functions for assessing predictive accuracy

Main functions:

• fx_scramble: Create df with scrambled group assignment
• fx_partition: List of partitions to apply to data frame
• fx_sample: Create train/test sub data frames
• fx_model: Train/test model on sub data frames
• fx_modelPerf: Confusion matrix and model performance metrics
• fx_rocCompute: Compute AUC of ROC
• fx_perm: Derive null distribution
• fx_boot: Bootstrap confidence intervals
• fx_permPerf: Estimate p-values, organize null distributions
• fx_bootPerf: Estimate CIs and p-values from bootstrap distributions
• fx_plot: Plot observed vs. null/bootstrap distributions
• fx_rocPlot: Estimate and plot ROC
• fx_outFile: Handles specified output files
• fx_summary: Produces .txt summary of model info (incomplete)

Installation

You can download the development version from GitHub using:

library(remotes)
install_github("fishpm/nruPredict")