supportHudgens: Determine the support of the NPMLE for interval censored...

In icmstate: Interval Censored Multi-State Models

Determine the support of the NPMLE for interval censored data.

Description

Given censoring/truncation intervals, find the maxcliques and determine the support of the interval censored problem.

Usage

supportHudgens(intervals, reduction = TRUE, existence = FALSE)

Arguments

intervals	A data.frame with 3 columns containing half-open intervals (left open, right closed) and an indicator whether the interval results from a censored transition or truncation: L: Left side of interval; R: Right side of interval; cens: Indicator whether interval resulted from interval censoring or left truncation (1 = censoring, 0 = truncation); id: (optional) Identifier for the observation this interval belongs to (numeric/integer). Only required if existence = TRUE; Note that the truncation intervals need to be in the form (N, Inf] with N a numeric value.
reduction	Should the support be reduced using Lemma 3 from Hudgens (2005)? This requires checking an extra condition. Default is TRUE.
existence	Should the existence of the NPMLE be checked using Theorem 1/Lemma 4 from Hudgens (2005)? Requires id to be present in intervals. Default is FALSE.

Value

• graph: An igraph object representing the censoring/truncation intervals

• support: Support estimated from the censoring intervals

• dir_graph: A directed igraph object used to determine whether the NPMLE exists in the presence of left-truncation.

• exist_mle: Logical output indicating whether the NPMLE exists.

References

Michael G. Hudgens, On Nonparametric Maximum Likelihood Estimation with Interval Censoring and Left Truncation, Journal of the Royal Statistical Society Series B: Statistical Methodology, Volume 67, Issue 4, September 2005, Pages 573-587, \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1111/j.1467-9868.2005.00516.x")}

icmstate documentation built on April 3, 2025, 8:06 p.m.