Reference

Package

precrec

precrec: A package for computing accurate ROC and Precision-Recall curves

Main

evalmod()

Evaluate models and calculate performance evaluation measures

Data preparation

mmdata()

Reformat input data for performance evaluation calculation

join_scores()

Join scores of multiple models into a list

join_labels()

Join observed labels of multiple test datasets into a list

format_nfold()

Create n-fold cross validation dataset from data frame

create_sim_samples()

Create random samples for simulations

Visualization

plot(<sscurves>) plot(<mscurves>) plot(<smcurves>) plot(<mmcurves>) plot(<sspoints>) plot(<mspoints>) plot(<smpoints>) plot(<mmpoints>)

Plot performance evaluation measures

autoplot(<sscurves>) autoplot(<mscurves>) autoplot(<smcurves>) autoplot(<mmcurves>) autoplot(<sspoints>) autoplot(<mspoints>) autoplot(<smpoints>) autoplot(<mmpoints>)

Plot performance evaluation measures with ggplot2

fortify(<sscurves>) fortify(<mscurves>) fortify(<smcurves>) fortify(<mmcurves>) fortify(<sspoints>) fortify(<mspoints>) fortify(<smpoints>) fortify(<mmpoints>)

Convert a curves and points object to a data frame for ggplot2

Data retrieval

as.data.frame(<sscurves>) as.data.frame(<mscurves>) as.data.frame(<smcurves>) as.data.frame(<mmcurves>) as.data.frame(<sspoints>) as.data.frame(<mspoints>) as.data.frame(<smpoints>) as.data.frame(<mmpoints>) as.data.frame(<aucroc>)

Convert a curves and points object to a data frame

auc()

Retrieve a data frame of AUC scores

pauc()

Retrieve a data frame of pAUC scores

Partical AUC and partial curve

part()

Calculate partial AUCs

Confidence interval of AUC scores

auc_ci()

Calculate CIs of ROC and precision-recall AUCs

Datasets

P10N10, B500, B1000, IB500, IB1000 and M2N50F5

P10N10

A small example dataset with several tied scores.

B500

Balanced data with 500 positives and 500 negatives.

B1000

Balanced data with 1000 positives and 1000 negatives.

IB500

Imbalanced data with 500 positives and 5000 negatives.

IB1000

Imbalanced data with 1000 positives and 10000 negatives.

M2N50F5

5-fold cross validation sample.