On Algorithms for the Nonparametric Maximum Likelihood Estimator of the Failure Function With Censored Data

Ying Zhang, Mortaza Jamshidian

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

In this article, we study algorithms for computing the nonparametric maximum likelihood estimator (NPMLE) of the failure function with two types of censored data: doubly censored data and (type 2) interval-censored data. We consider two projection methods, namely the iterative convex minorant algorithm (ICM) and a generalization of the Rosen algorithm (GR) and compare these methods to the well-known EM algorithm. The comparison conducted via simulation studies shows that the hybrid algorithms that alternately use the EM and OR for doubly censored data or, alternately, use the EM and ICM for (type 2) interval-censored data appear to be much more efficient than the EM, especially in large sample situation.

Original languageEnglish (US)
Pages (from-to)123-140
Number of pages18
JournalJournal of Computational and Graphical Statistics
Volume13
Issue number1
DOIs
StatePublished - Mar 1 2004
Externally publishedYes

Fingerprint

Nonparametric Maximum Likelihood Estimator
Censored Data
Maximum likelihood
Doubly Censored Data
Interval-censored Data
EM Algorithm
Hybrid Algorithm
Projection Method
Simulation Study
Maximum likelihood estimator
Censored data
Computing

Keywords

  • Double censoring
  • EM algorithm
  • Gradient projection algorithm
  • Interval censoring
  • Iterative convex minorant algorithm
  • Rosen method

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability
  • Computational Mathematics

Cite this

On Algorithms for the Nonparametric Maximum Likelihood Estimator of the Failure Function With Censored Data. / Zhang, Ying; Jamshidian, Mortaza.

In: Journal of Computational and Graphical Statistics, Vol. 13, No. 1, 01.03.2004, p. 123-140.

Research output: Contribution to journalArticle

@article{c7572f5a9d654d93b740986ce1f75a9d,
title = "On Algorithms for the Nonparametric Maximum Likelihood Estimator of the Failure Function With Censored Data",
abstract = "In this article, we study algorithms for computing the nonparametric maximum likelihood estimator (NPMLE) of the failure function with two types of censored data: doubly censored data and (type 2) interval-censored data. We consider two projection methods, namely the iterative convex minorant algorithm (ICM) and a generalization of the Rosen algorithm (GR) and compare these methods to the well-known EM algorithm. The comparison conducted via simulation studies shows that the hybrid algorithms that alternately use the EM and OR for doubly censored data or, alternately, use the EM and ICM for (type 2) interval-censored data appear to be much more efficient than the EM, especially in large sample situation.",
keywords = "Double censoring, EM algorithm, Gradient projection algorithm, Interval censoring, Iterative convex minorant algorithm, Rosen method",
author = "Ying Zhang and Mortaza Jamshidian",
year = "2004",
month = "3",
day = "1",
doi = "10.1198/1061860043038",
language = "English (US)",
volume = "13",
pages = "123--140",
journal = "Journal of Computational and Graphical Statistics",
issn = "1061-8600",
publisher = "American Statistical Association",
number = "1",

}

TY - JOUR

T1 - On Algorithms for the Nonparametric Maximum Likelihood Estimator of the Failure Function With Censored Data

AU - Zhang, Ying

AU - Jamshidian, Mortaza

PY - 2004/3/1

Y1 - 2004/3/1

N2 - In this article, we study algorithms for computing the nonparametric maximum likelihood estimator (NPMLE) of the failure function with two types of censored data: doubly censored data and (type 2) interval-censored data. We consider two projection methods, namely the iterative convex minorant algorithm (ICM) and a generalization of the Rosen algorithm (GR) and compare these methods to the well-known EM algorithm. The comparison conducted via simulation studies shows that the hybrid algorithms that alternately use the EM and OR for doubly censored data or, alternately, use the EM and ICM for (type 2) interval-censored data appear to be much more efficient than the EM, especially in large sample situation.

AB - In this article, we study algorithms for computing the nonparametric maximum likelihood estimator (NPMLE) of the failure function with two types of censored data: doubly censored data and (type 2) interval-censored data. We consider two projection methods, namely the iterative convex minorant algorithm (ICM) and a generalization of the Rosen algorithm (GR) and compare these methods to the well-known EM algorithm. The comparison conducted via simulation studies shows that the hybrid algorithms that alternately use the EM and OR for doubly censored data or, alternately, use the EM and ICM for (type 2) interval-censored data appear to be much more efficient than the EM, especially in large sample situation.

KW - Double censoring

KW - EM algorithm

KW - Gradient projection algorithm

KW - Interval censoring

KW - Iterative convex minorant algorithm

KW - Rosen method

UR - http://www.scopus.com/inward/record.url?scp=1842434519&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=1842434519&partnerID=8YFLogxK

U2 - 10.1198/1061860043038

DO - 10.1198/1061860043038

M3 - Article

VL - 13

SP - 123

EP - 140

JO - Journal of Computational and Graphical Statistics

JF - Journal of Computational and Graphical Statistics

SN - 1061-8600

IS - 1

ER -