This package is for calculating the estimates for the logistic
regression coefficients in the models presented in the paper
Estimation And Inference For Logistic Regression With Covariate
Missclassification And Measurement Error, In Main Study/Validation
Study Designs.
Donna Spiegelman, Harvard School Of Public Health,
Departments Of Epidemiology And Biostatistics.
stdls@channing.harvard.edu
Bernard Rosner, The Channing Laboratory,
Brigham and Women's Hospital.
Roger Logan, Harvard School of Public Health,
Department of Epidemiology.
More information on the program is contained in the file betacomp.f.
This package should include the files:
files:
betacomp.f : Main program for calculating estimates of the betas.
likroutines.f : Subroutines involving the likelihood and derivatives.
logit2.f : Logistic regression subroutine.
blin.f : Regression Calibration subroutine.
mvlm2.f : multivariate linear regression subroutine.
matrix.f : Linear algegra routines needed by program.
optimize.f : Subroutines needed for dmnhb optimizer.
misc.f : Subroutines needed for outputing of results.
Makefile : Makefile for building executable: betacomp.
mainstudy.dat : Sample main study data file.
validstudy.dat: Sample validation study data file.
sample.out : Sample output.
manifest : list of included files
To create the executable use the command
make betacomp
This will create the main program called betacomp. For further information on
this program contact Donna Spiegelman at stdls@channing.harvard.edu
The optimizer is dmnhb which was obtained from Netlib. There is some
documentation included in the file optimize.f. Additional documentation
can be obtained off the web from Netlib in the file
http://www.netlib.org/port/dmng.f
It is not difficult to change the optimizer to use routine provided by IMSL. The
subroutines liglik24, score24a, and hess24a are set up for IMSL routines.