PerfCurveBands: Construct a confidence band for a recall or precision curve

In jrash/chemmodlab: A Cheminformatics Modeling Laboratory for Fitting and Assessing Machine Learning Models

Construct a confidence band for a recall or precision curve

Description

PerfCurveBands takes a pair of score and activity vectors as input. A performance curve and confidence band is created for the selected testing fractions.

Usage

PerfCurveBands(
  S,
  X,
  r,
  metric = "rec",
  type = "band",
  method = "sup-t",
  plus = T,
  conf.level = 0.95,
  boot.rep = 100,
  mc.rep = 1e+05,
  myseed = 111,
  h = NULL
)

Arguments

S	a vector of scores.
X	a vector of activities.
r	a vector of testing fractions.
metric	the performance curve to use. Options are recall ("rec") and precision ("prec").
type	specifies whether a point-wise confidence interval ("pointwise") or a confidence band ("band") should be constructed.
method	the method to use. Point-wise confidence interval options are "binomial", "JZ", "bootstrap". Confidence band options are "sup-t", "theta-proj".
plus	should plus correction be used or not?
conf.level	the confidence level for the bands.
boot.rep	the number of replicates to use for the bootstrap method.
mc.rep	the number of Monte Carlo replicates to use for the sup-t method.
myseed	the random seed.
h	the bandwidth for the local regression estimator of Lambda. If NULL, uses the default plugin estimator.

jrash/chemmodlab documentation built on May 18, 2023, 8:42 p.m.