Learn R Programming

bootLR (version 1.0.2)

Bootstrapped Confidence Intervals for (Negative) Likelihood Ratio Tests

Description

Computes appropriate confidence intervals for the likelihood ratio tests commonly used in medicine/epidemiology, using the method of Marill et al. (2015) . It is particularly useful when the sensitivity or specificity in the sample is 100%. Note that this does not perform the test on nested models--for that, see 'epicalc::lrtest'.

Copy Link

Version

Install

install.packages('bootLR')

Monthly Downloads

182

Version

1.0.2

License

LGPL-2.1

Maintainer

Last Published

February 1st, 2019

Functions in bootLR (1.0.2)

BayesianLR.test

Compute the (positive/negative) likelihood ratio with appropriate, bootstrapped confidence intervals

Compute values and confidence intervals for sensitivity, specificity, positive likelihood ratio, negative likelihood ratio for a single 2x2 table

confusionStatistics

Compute sensitivity, specificity, positive likelihood ratio, negative likelihood ratio for a single 2x2 table

Internal function to draw a set of sensitivities or specificities This is intended for the case where testPos == totalDzPos or testNeg == totalDzNeg.

medianConsistentlyOne

Find the lowest population probability whose median is consistently one This is the lowest estimate for Sens that is consistently (over 5 runs) most likely to yield a sample estimate that is all 1's (e.g. 100/100, etc.).

Prints results from diagCI As is typical for R, this is run automatically when you type in an object name, and is typically not run directly by the end-user.

Prints results from the BayesianLR.test As is typical for R, this is run automatically when you type in an object name, and is typically not run directly by the end-user.

run.BayesianLR.test

The actual function that runs the test. BayesianLR.test is a wrapper that runs this with ever-looser tolerances.

sequentialGridSearch

Optimize a function returning a single numeric value subject to a boolean constraint Utilizes a naive recursive grid search.

Internal function to analyze LR bootstrap finding median, and standard and BCa percentile 95 To obtain bca CI on a non-boot result, use a dummy boot. and replace t and t0 with the results of interest.