ROC Curve Analysis

Function for ROC (Receiver Operating Characteristic) Curve Analysis. Calculates sensitivity, specificity, AUC, Youden index, and more.

Usage

roc(
  data,
  classvar,
  classpos,
  testvar,
  direction = "greatpos",
  ci = TRUE,
  cimethod = "delong",
  youden = TRUE,
  optimcrit = TRUE,
  pp = FALSE,
  pprob = 0.5,
  costratioFP = 1,
  coords = TRUE,
  plotroc = TRUE,
  plotci = FALSE,
  plotbars = FALSE,
  plotprev = FALSE,
  plotidr = TRUE
)

Arguments

data

The data as a data frame.

classvar

A string naming the variable from data that contains the actual status (gold standard / diagnosis), must be a factor with two levels.

classpos

The level of the classification variable that represents the positive status.

testvar

A string naming the variable from data that contains the test result or measurement, must be numeric.

direction

Specify whether greater values of the test variable indicate a positive test (default) or if lower values indicate a positive test. Note: AUC will always be greater than 0.5, even if it means inverting the predictor internally.

ci

Boolean selection whether to show 95\ Default is 'true'.

cimethod

Method for calculating confidence intervals for the AUC. Default is 'delong'.

youden

Boolean selection whether to calculate Youden index (J = sensitivity + specificity - 1). Default is 'true'.

optimcrit

Boolean selection whether to calculate the optimal criterion value. Default is 'true'.

pp

Boolean selection whether to use a specified disease prevalence for predictive values. Default is 'false'.

pprob

Prior probability (disease prevalence in the population). Requires a value between 0.001 and 0.999, default 0.500.

costratioFP

Cost ratio of false positive to false negative decisions. Default is 1.000 (equal costs).

coords

Boolean selection whether to show the full list of coordinates of the ROC curve. This includes all possible thresholds with sensitivity, specificity, predictive values, and likelihood ratios. Default is 'true'.

plotroc

Boolean selection whether to create a ROC curve plot. Default is 'true'.

plotci

Boolean selection whether to show 95\ the ROC curve. Default is 'false'.

plotbars

Boolean selection whether to create a plot of sensitivity and specificity versus criterion values. Default is 'false'.

plotprev

Boolean selection whether to create a plot of positive and negative predictive values versus disease prevalence. Default is 'false'.

plotidr

Boolean selection whether to create an interactive dot diagram Default is 'true'.

Value

A results object containing:

results$summary					a table
results$optimal					a table
results$coords					a table
results$plotroc					an image
results$plotbars					an image
results$plotprev					an image
results$plotidr					an image

Tables can be converted to data frames with asDF or as.data.frame. For example:

results$summary$asDF

as.data.frame(results$summary)

Examples

# \donttest{
# example will be added
# }