------------------------- ANNOUNCE ALL -------------------------
This file contains a compendium of all code available from the Section
of Computer Science, Department of Biomathematics, University of Texas
M. D. Anderson Hospital. This code can be obtained by anonymous ftp
to odin.mda.uth.tmc.edu (129.106.3.17); the current file is
./pub/index. More information on obtaining the code can be found at
the bottom of this listing.
Computer code written by members of the Department of Biomathematics,
the University of Texas M. D. Anderson Cancer Center is now available.
The code can be freely copied and used (shareware distribution is
encouraged) although the authors retain copyright for the University
of Texas in order to control possible commercial incorporation. The
last section of this message contains instructions on how to obtain
it.
======================================================================
ACCFLF
Accelerated Failure Time Model with Log-F Distribution
Barry W. Brown
F. Martin Spears
Lawrence Levy
James Lovato
Kathy Russell
CAPABILITIES
This program fits the accelerated failure model to data with or
without covariates. Its capabilities include fitting the general
model or any submodel with p and q fixed. It also can automatically
fit to data a set of models specified by a rectangular grid in (p,q)
coordinates or a set of named models. Several options are provided to
add covariates to an existing model. Finally, ACCFLF will calculate
time to event probabilities for accelerated failure models using
either times from the current data or a user supplied list.
Reference
Kalbfleish, J.D. and Prentice, R. L. (1980) The Statistical Analysis
of Failure Time Data. New York, John Wiley and Sons.
FILES
./pub/accflf/accflf-2.0-tar.Z: Fortran 77 source and documentation.
./pub/accflf/accflf20.exe: self-extracting MSDOS executable and
documentation.
./pub/accflf/paper.tex.Z: TeX draft of a paper describing the
calculations of the likelihood and some derivatives.
./pub/accflf/paper.ps: Postscript version of this paper.
======================================================================
BUSSPA
Binomial Unequal Sample Size Power and Analysis
BUSSPA performs computations for testing for differences in binomial
probabilities P1 and P2 in two groups. The number of trials may
differ in the two groups.
CAPABILITIES
Calculates exact significance levels.
Performs calculations to plan a trial.
FILES
/pub/source/busspa.tar.Z: Fortan 77 source and documentation.
/pub/msdos/busspapk.exebusspapk.exe: Self-extracting MSDOS file
containing executable and documentation
/pub/mac/busspa.sea.hqx: Macintosh self-extracting file
containing executable and documentation
======================================================================
DCDFLIB
Library of Routines for Cumulative Distribution
Functions, Inverses, and Other Parameters
C and Fortran
Barry W. Brown
James Lovato
Kathy Russell
LATEST VERSION: 1.0
CAPABILITIES
DCDFLIB (Double precision Cumulative Distribution Function LIBrary) is
a collection of routines that calculate cumulative distribution
functions, inverses, and parameters for common statistical
distributions. DCDFLIB uses published algorithms where available, and
literature citations are included in the documentation.
Values associated with a statistical distribution include X, the upper
limit of integration of the density, P, the cumulative distribution
function evaluated at X, and auxillary parameters such as degrees of
freedom. Given all but one such value, a routine in cdflib will
calculate the one value. (CAVEAT: For the central and noncentral F,
the cdf is sometimes not monotone in the degrees of freedom. For
these parameters, there can be two answers. DCDFLIB finds an
arbitrary one of the two in this case.
Routines are provided for the following distributions.
(1) Beta
(2) Binomial
(3) Chi-square
(4) Noncentral Chi-square
(5) F
(6) Noncentral F
(7) Gamma
(8) Negative Binomial
(9) Normal
(10) Poisson
(11) Student's t
FILES
dcdflib.c-1.0-tar.Z: Source in Ansi C plus documentation.
dcdflib.f-1.0-tar.Z: Source in Fortran 77 plus documentation.
dcdflib.kr.c-1.0-tar.Z: Source in K&R C (old C) plus documentation.
MSDOS versions are self-extracting files packaged with a licensed
copy of PKZIP.
./msdos/dcdc10.exe: Source in ANSI C
./msdos/dcdf10.exe: Source in F77
======================================================================
DSTATTAB
Calculates Statistical Cumulative Distribution Functions (P-Values Their Inverses, and Parameters of the Distributions.
Barry W. Brown
David Gutierrez
James Lovato
Marty Spears
LATEST VERSION 1.0
WHAT'S NEW? (Compared to STATTAB.) Uses double precision library,
DCDFLIB, and so is more accurate. Packaging changed to allow entry of
user's choice of X or 1-X where appropriate and display of both. This
allows calculations of p-values quite near one and accurate cdf values
that are also close to 1.
CAPABILITIES
DSTATTAB calculates the cumulative distribution functions, inverses,
and values of parameters for the statistical distributions listed
below.
1 - Incomplete Beta
2 - Binomial
3 - Negative Binomial
4 - Chi-square
5 - Non-central Chi-square
6 - F
7 - Non-central F
8 - Incomplete Gamma
9 - Normal
10 - Poisson
11 - T
It can also produce tables of any one calculated item as a function of
a specified list of values of another item.
FILES
./pub/source/dstattab-1.0-tar.Z: F77 source plus docs
./pub/msdos/dstatt10.exe: Self extracting PC executable plus docs
./pub/mac/dstattab-1.0-sea.hqx: Macintosh executable plus docs
======================================================================
EXTSIG
Exact Significance Level for 2 X 2 Tables
Barry W. Brown
David Gutierrez
F. Martin Spears
CAPABILITIES
EXTSIG computes exact significance levels and the usual chi-square
statistic for testing the difference in two binomial proportions.
FILES
./source/extsig.tar.Z: Source in Fortran 77 plus documentation
./pub/msdos/extsigpk.exe: MSDOS self-extracting file containing
executable and documentation
./pub/mac/extsig.sea.hqx: Macintosh self-extracting file containing
executable and documentation
======================================================================
KSB1CI
K-Stage Binomial One-Group Confidence Intervals
Barry W. Brown
CAPABILITIES
Finds confidence intervals for single- or multiple-stage trials
involving a single group which have a binomial outcome. Calculated by
a minor modification of the Clopper and Pearson method to k-stage use.
FILES
./pub/source/ksb1ci.tar.Z: Source in Fortran 77 plus documentation.
./pub/msdos/ksb1cipk.exe: Self-extracting MSDOS file containing
executable plus documentation
./pub/mac/ksb1ci.sea.hqx: Self-extracting file for Macintosh
containing executable plus documentation.
======================================================================
KSBIN1
K-Stage BINomial 1 Sample
Barry W. Brown
CAPABILITIES
Planning of exact single- or multiple-stage experiments on a single
group with binomial outcome of H0:P=P0 versus HA:P=PA. At each stage
the trial can be terminated because of early proof of HA or because
the outcome is such that it is unlikely that proof for HA will be
obtained even if HA is correct.
FILES
./pub/source/ksbin1.tar.Z: Source in Fortran 77 plus documentation.
./pub/msdos/ksbin1pk.exe: Self-extracting MSDOS file containing
executable plus documentation.
./pub/mac/ksbin1.sea.hqx: Self-extracting Macintosh file containing
executable plus documentation
======================================================================
MATHLIB
(Source only -- no MSDOS or Mac executables)
MATHLIB contains the mathematical and statistical routines used in the
other applications. Typical functions are cumulative and inverse
cumulative distribution functions, finding a zero of a monotone
function. Where possible, MATHLIB routines were obtained from
published sources, such as the ACM Transactions on Mathematical
Software.
---------- MATHLIB NOTE ----------
November 11, 1993
The current version of MATHLIB is obsolete, but I am leaving it where
it is for the moment. Most of the functionality of MATHLIB is found
in CDFLIB and RANLIB and the routines in those libraries are generally
better than the ones in MATHLIB. A much smaller (and more eclectic)
version of MATHLIB will be posted when we get around to it.
======================================================================
RANLIB
Library of Routines for Random Number Generation
(C and Fortran)
Barry W. Brown
James Lovato
Kathy Russell
LATEST VERSION: 1.1
WHAT'S NEW? Bugs fixed (primarily SAVE statements in Fortran
generator for Gamma distribution), error messages go to stderr (C
version). Thanks to users who caught this. Added generators for
multinomial and negative binomial distributions.
CAPABILITIES
The bottom level routines provide 32 virtual random number generators.
Each generator can provide 1,048,576 blocks of numbers, and each block
is of length 1,073,741,824. Any generator can be set to the beginning
or end of the current block or to its starting value. Packaging is
provided so that if these capabilities are not needed, there is a
single generator with period 2.3 X 10^18.
Using this base, routines are provided that return:
(1) Beta random deviates
(2) Chi-square random deviates
(3) Exponential random deviates
(4) F random deviates
(5) Gamma random deviates
(6) Multivariate normal random deviates (mean and covariance
matrix specified)
(7) Noncentral chi-square random deviates
(8) Noncentral F random deviates
(9) Univariate normal random deviates
(10) Random permutations of an integer array
(11) Real uniform random deviates between specified limits
(12) Binomial random deviates
(13) Negative Binomial random deviates
(14) Multinomial random deviates
(15) Poisson random deviates
(16) Integer uniform deviates between specified limits
(17) Seeds for the random number generator calculated from a
character string
FILES
ranlib.c-1.1-tar.Z: Ansi C version with documentation
ranlib.f-1.1-tar.Z: Fortran 77 version with documentation
MSDOS versions are self-extracting files packaged with a licensed
copy of PKZIP.
./msdos/ranc10.exe: Source in ANSI C
./msdos/ranf10.exe: Source in F77
======================================================================
RANLIST
Randomization Lists
Barry W. Brown
James Lovato
Benjamen Rundell
CAPABILITIES
Produces randomization lists.
Features both restricted and unrestricted randomization. When
restricted randomization is used, fixed or random balance points may
be chosen.
Offers a primitive form of interactive randomization.
FILES
./source/ranlist-1.1-tar.Z: Fortran 77 source and documentation.
./pub/msdos/ranlst11.exe: Self-extracting MSDOS file containing
executable and documentation
./pub/mac/ranlist-1.1-sea.hqx: Macintosh executable plus docs
======================================================================
STPLAN 4.0
Study Planning
Barry W. Brown
Christopher Brauner
Albert Chan
David Gutierrez
Jay Herson
James Lovato
James Polsley
CAPABILITIES
STPLAN performs power, sample size, and related calculations needed to
plan studies. Parameters of a test include those necessary to
describe the distributions under the null and alternative hypotheses,
sample sizes, significance levels, and power. Given values for all
but one parameter, STPLAN will calculate the value of that one
parameter.
The following testing situations are covered by STPLAN:
Studies with Outcomes Distributed According to the Binomial Distribution
One-sample test
Two-sample test
Two-sample above and below the median
Fisher's exact
Matched pairs test for equality of proportions
Conservative versus expensive treatment
Two-sample test in which one group is a historic control
K-sample
Case-control
Sample size correction for loss of subjects over time
Studies with Outcomes Distributed According to the Poisson Distribution
One-sample test
Two-sample test
Studies with Outcomes Distributed According to the Normal Distribution
One-sample test
Two-sample test, equal variances
Two-sample test, unequal variances
Two-sample log-normal
Studies with Outcomes Distributed According to the
Exponential Distribution
One-sample test
Two-sample test
Studies whose Outcome is a Set of Survival Times.
Randomized clinical trial for exponential survival times
Comparison with a historic control - exponential survival
Randomized clinical trial - two different hazard rates
depending on time
Studies whose Outcome is the Correlation Coefficient
One-sample test
FILES
./pub/source/stplan-4.0-tar.Z: Fortran source plus documentation
./pub/msdos/stplan40.exe: Self extracting file of MSDOS
executable plus documentation
./pub/mac/stplan-4.0-sea.hqx: Macintosh executable pus documentaion
======================================================================
SURVAN
Survival Analysis
Barry W. Brown
James Lovato
M. Elizabeth Rozell
David Tuttle
CAPABILITIES
Survival calculations including Kaplan-Meier estimation, log-rank and
Gehan-Breslow tests, proportional hazards and logistic regression
models.
FILES
./pub/source/survan.tar.Z: Fortran source plus documentation
./pub/msdos/survanok.exe: Self extracting file of MSDOS
executable plus documentation
./pub/mac/survan.sea.hqx: Self extracting file for Macintosh
executable plus documentation
======================================================================
WINDOWS
Windowed Estimates
Barry W. Brown
David Gutierrez
James Lovato
F. Martin Spears
Patrick Tibbits
CAPABILITIES
WINDOWS provides a variety of window/kernel/moving-average estimators.
Both fixed width and fixed number of nearest neighbor windowing is
provided. Weighting of points may be done either with boxcar or
biquartic weights. Estimates available include weighted mean, a
linear or quadratic fit, maximum or minimum value, a specified
percentile, the standard deviation, and the derivative of y with
respect to x. Cross validation is also included.
FILES
./pub/source/windows.tar.Z: F77 source plus documentation
./pub/msdos/windowpk.exe: Self extracting file for MSDOS
containing executable plus
documentation
./pub/mac/windows.sea.hqx: Self extracting file for Macintosh
contains executable plus docs
======================================================================
S-Stuff
This collection consists of source code and documentation for the S
language. We submit our S developments to statlib and generally
delete them from this local ftp site when they are placed there;
consequently, this is not a complete collection of our developments.
The files in this collection are on directory ./pub/S.
Each file is a compressed shell archive (shar file). To extract a
file called blip.shar.Z on Unix, place it in a directory by itself and
type
zcat blip.shar.Z | /bin/sh
blip
Draws one-dimensional distribution plots using a number of fancy
options. Blip can be used to construct standard or customized
distribution plots.
bphase2
This directory contains Splus routines to create Bayesian Phase II
monitoring boundaries. Due to peculuarities in the way files are
handled these routines will work only with Splus for Unix.
dgay
Dgay.f is a file of the FORTRAN routines written by David Gay. These
routines have been modified so that they can be loaded in S. (These
routines are provided as part of Splus with a better interface, so
Splus users shouldn't bother.)
integrate
Includes routines to numerically integrate a function of one or more
variables. Includes FORTRAN and C code.
invmf
Routine to numerically invert a monotone function. Includes FORTRAN
and C code.
stukel
Generalized Logistic Regression, using the Stukel algorithm. A
parametric family of link functions generalizes logistic regression.
trax
Trax is an S function which transforms the x and/or y axis of a plot
by an arbitrary function. It is a generalization of the log plot,
useful, for example, in providing plots on a logit scale.
======================================================================
HOW TO GET IT
I. This file can be obtained by the command
finger software@odin.mda.uth.tmc.edu
II. Accessing us by WWW
For general information on the University of Texas M. D. Anderson
Cancer Center:
http://utmdacc.mda.uth.tmc.edu
For information on the Department of Biomathematics of the above
institution:
http://odin.mda.uth.tmc.edu
For specific information on available code:
http://odin.mda.uth.tmc.edu/anonftp/
III. By anonymous ftp to us
The code is available by anonymous ftp to odin.mda.uth.tmc.edu
(129.106.3.17).
Directory source contains s ource code as tar.Z files. The source
includes Mac, MSDOS, UNIX, and Vax specific code where it is needed.
The code is written in portable Fortran 77 and moving it to other
machines is generally easy. To port it, look at the comments in the
machine specific code and write the same function for the new machine.
The remainder of the code should compile and run.
Directory 'msdos' contains MSDOS executables (and documentation);
directory 'mac' contains the same for Macintosh machines. The
executables for both machines are in self-extracting compressed files.
IV.. From statlib
The source of everything that we make public is posted to statlib
(eventually). The S collection of statlib includes postings by us
that are not available locally. Much of the remainder of the
collection is in "utexas", although some is in collection "general".
The statlib WWW address is:
http://lib.stat.cmu.edu/
Statlib can also be reached by email at
statlib@temper.stat.cmu.edu.
Typical messages are:
send help
(A help file describing statlib is returned by email.)
send index
(An overall index to the collection is returned.
send index from general
send index from utexas
send index from S
(An index to collections general, utexas and S are returned.
Barry W. Brown, Professor
of Biomathematics
Section of Computer Science
Department of Biomathematics,
Box 237
University of Texas M. D.
Anderson Hospital
1515 Holcombe Blvd
Houston, TX 77030
bwb@odin.mda.uth.tmc.edu
internet address is (129.106.3.17)