An Algorithm for the Computation of Generalized Likelihood or Self-Critical Estimators for Binary Data.
Abstract
This paper describes the computational algorithms for computing generalized likelihood estimators for parametric proportional hazards models. Subroutine BINARY is an implementation of the self-critical estimation procedure of Paulson, Presser and Lawrence (1983). The logistic, Gaussian or Type I extreme value distribution may be selected as tolerance distribution. Estimates are expressed in location-scale form on entry and exit, but results may be printed out in regression form, location-scale form, or in both forms. It is possible to hold location parameters constant during the estimation procedure. Estimation is accomplished by a Newton-Raphson method. Keywords: FORTRAN; Mathematical Models.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1986
- Accession Number
- ADA178545
Entities
People
- A. S. Paulson
- T. A. Delahanty