Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

PREG: a computer program for Poisson regression analysis

Technical Report ·
OSTI ID:6519145

PREG is a FORTRAN program that computes maximum likelihood estimates for a regression analysis when the dependent variable is a count that follows the Poisson distribution. This documentation contains a detailed description of the algorithm for maximum likelihood estimation under the Poisson regression model (subroutine PRMLE), three test problems with complete output (Appendix A), and a complete listing of the entire FORTRAN program, PREG-Version 5, Release 1 (Appendix B). The three test problems illustrate Poisson regression with (i) a linear regression model, (ii) a log-linear regression model, and (iii) a nonlinear regression model. This program was developed using FORTRAN-10 on a DEC system-10 computer and has been compiled and executed on an IBM 360/91 under OS and an IBM 3033 under VMS without any changes in the source code. ANSI standard FORTRAN was used in all program units except subroutine PRDATA, which can be easily modified to suit individual user requirements.

Research Organization:
Oak Ridge Associated Universities, Inc., TN (USA)
DOE Contract Number:
AC05-76OR00033
OSTI ID:
6519145
Report Number(s):
ORAU-178
Country of Publication:
United States
Language:
English