EM.perept: EM.perept

In alex-kalinka-cruk/perept: Performance estimation using repeated tests.

Description

Expectation Maximisation (EM) algorithm to estimate false-positive and false-negative rates of a test when the true response is unknown and repeated tests are available for a set of test subjects (e.g. genes).

Usage

EM.perept(n_pos, N, delta_pos = NULL, maxiter = 1e+05, tol = 1e-06)

Arguments

n_pos	Vector of integers specifying the number of tests that produced a positive result for each subject.
N	Integer value giving the total number of repeated tests per subject.
delta_pos	A vector of real numbers (in the range [0,1]) giving initial estimates of the probability of each subject being positive - must be the same length as 'n_pos'. If 'NULL', then 'n_pos'/'N' is used. Defaults to 'NULL'.
maxiter	A positive integer specifying the maximum allowable number of iterations without convergence occurring. Defaults to 1e5.
tol	A real number specifying the maximum ODL difference between successive steps of the EM algorithm below which convergence occurs. Defaults to 1e-6.

Value

An object of class 'EM.perept', which is a list with the following elements:

References

Jakobsdottir and Weeks (2007). Estimating prevalence, false-positive rate, and false-negative rate with use of repeated testing when true responses are unknown. Am J Hum Genet 81:1111-1113.

alex-kalinka-cruk/perept documentation built on Jan. 11, 2020, 3:46 p.m.