Evgenii B. Rudnyi
Blog http://blog.rudnyi.ru
Name Last modified Size
Parent Directory 06-Apr-2011 04:18 -
dat/ 06-Apr-2011 04:18 -
out/ 06-Apr-2011 04:18 -
outold/ 06-Apr-2011 04:18 -
COPYING 06-Jan-2005 15:45 18k
FREE_LICENSE 06-Jan-2005 15:45 1k
generate.cpp 05-Jan-2005 00:33 8k
line.cfg 04-Jan-2005 21:38 1k
linear.cpp 05-Jan-2005 00:33 2k
makefile 04-Jan-2005 21:31 1k
oneway.cfg 04-Jan-2005 21:36 1k
readme.txt 21-Jul-1994 23:20 17k
Copyright (C) 1994-1995 Evgenii Rudnyi, http://Evgenii.Rudnyi.Ru/
This software is a copyrighted work licensed under the terms,
described in the file "FREE_LICENSE".
Note that the license written as the code output concerns the binary
(the same for readme.txt). The source code is covered by
"FREE_LICENSE".
Use at your own risk.
This file is written in 2005.
Software linear and generate
============================
This is software to accompany my paper [1].
readme.txt is the old manual written in 1994. It looks like that it is
still correct but I have not checked it completely.
I have fixed files and now they can be compiled with gcc 3.3 on my
Mac. Mostly this was necessary for generate.cpp as it used some
Borland specific calls. It still requires GGNML from IMSL to generate
random numbers according to the Gauss distribusion.
linear.cpp is a simple driver designed to work with my varcomp
library.
This example is also included in tdlib where the application is
defined in terms of datafiles only. With tdlib one does not need
linear.cpp.
Files
-----
readme.txt is the old manual.
generate.cpp is the code for generate.
oneway.cft is the sample configuration file for generate.
line.cfg is the sample configuration file for generate.
linear.cpp is the code for linear.
dat/ is the directory with the original datafiles for the paper.
dat/run is a script to run all the datafiles.
dat/oneway1 to dat/oneway4 are datafiles for the cases "a" to "d" in Table 1.
dat/line1 to dat/line4 are datafiles for the cases "a" to "d" in Table 2.
dat/oneway.set contains both ML and OLS settings for Table 1.
dat/outline.set contains both ML and OLS settings for Table 2.
out/ is the directory with the listings that I have received in 2005.
outold/ is the directory with the old listings from 1995.
The listings are quite similar (use diff -b to compare them). There is
a slight difference in convergence of ZXSSQ (it is normal for
numerics) and a slight difference in the output format. The results
are the same.
[1] E.B. Rudnyi. Statistical model of systematic errors: linear error
model. Chemometrics and Intelligent Laboratory Systems. 1996, V. 34, N
1, p. 41-54.