cal: Calibration Indices

In mattwarkentin/precogs: Risk Prediction Model Assessment

Description

Get summary calibration indices for binary outcomes and class probabilities. Optionally, get bootstrap-based confidence intervals using validation data.

Usage

cal_ind(
  data,
  y,
  p,
  val = NULL,
  conf.int = FALSE,
  level = 0.95,
  na.rm = FALSE,
  span = 0.75,
  ...
)

cal_ICI(p, p.calibrate)

cal_E50(p, p.calibrate)

cal_E90(p, p.calibrate)

cal_Emax(p, p.calibrate)

Arguments

data	Data frame.
y	Binary outcome.
p	Probabilites.
val	Data frame.
conf.int	Logical. Default is FALSE.
level	Confidence level. Default is 0.95.
na.rm	Logical. Default is FALSE.
span	Passed on to loess.
...	Arguments passed on to bootstraps.
p.calibrate	Calibrated predictions.

Value

A tibble with as many rows and calibration indices. The number of columns can vary based on whether confident intervals are requested.

Examples

dd <- data.frame(y = rbinom(50, 1, 0.5), p = runif(100, 0, 1))
cal_ind(dd, y, p)

mattwarkentin/precogs documentation built on Jan. 12, 2020, 6:24 p.m.