DirichletMultinomial (version 1.14.0)

roc: Summarize receiver-operator characteristics

Description

Returns a data.frame summarizing the cummulative true- and false-positive probabilities from expected and observed classifications.

Usage

roc(exp, obs, ...)

Arguments

exp

logical() vector of expected classifications to a particular group.

obs

Predicted probability of assignment to the group identified by TRUE values in exp. The length of exp and obs must be identical.

...

Additional arguments, available to methods.

Value

A data.frame with columns

TruePositive

Cummulative probability of correct assignment.

FalsePositive

Cummulative probability of incorrect assignment.

Examples

library(lattice)

## count matrix
fl <- system.file(package="DirichletMultinomial", "extdata",
                  "Twins.csv")
count <- t(as.matrix(read.csv(fl, row.names=1)))

## phenotype
fl <- system.file(package="DirichletMultinomial", "extdata",
                  "TwinStudy.t")
pheno0 <- scan(fl)
lvls <- c("Lean", "Obese", "Overwt")
pheno <- factor(lvls[pheno0 + 1], levels=lvls)
names(pheno) <- rownames(count)

## count data used for cross-validation, and cross-validation
count <- csubset(c("Lean", "Obese"), count, pheno)
data(bestgrp)

## true, false positives from single-group classifier
bst <- roc(pheno[rownames(count)] == "Obese",
           predict(bestgrp, count)[,"Obese"])
head(bst)

## lattice plot
xyplot(TruePostive ~ FalsePositive, bst, type="l",
       xlab="False Positive", ylab="True Positive")