EM {Icens}R Documentation

A function to compute the NPMLE of p based on the incidence matrix A.

Description

The incidence matrix, A is the m by n matrix that represents the data. There are m probabilities that must be estimated. The EM, or expectation maximization, method is applied to these data.

Usage

EM(A, pvec, maxiter=500, tol=1e-12)

Arguments

A The incidence matrix.
pvec The probability vector.
maxiter The maximum number of iterations.
tol The tolerance used to judge convergence.

Details

Lots.

Value

An object of class icsurv containing the following components:

pf The NPMLE of the probability vector.
numiter The number of iterations used.
converge A boolean indicating whether the algorithm converged.
intmap If present indicates the real representation of the support for the values in pf.

Author(s)

Alain Vandal and Robert Gentleman.

References

The EM algorithm applied to the maximal cliques of the intersection graph of the censored data. The empirical distribution function with arbitrarily grouped, censored and truncated data, B. W. Turnbull, 1976, JRSS;B.

See Also

VEM, ISDM, EMICM, PGM

Examples

    data(cosmesis)
    csub1 <- subset(cosmesis, subset= Trt==0, select=c(L,R))
    EM(csub1)
    data(pruitt)
    EM(pruitt)

[Package Icens version 0.5 Index]