For Immediate Release
Contact: Bob Simons, 800 728-9878,
info@cohort.com
MONTEREY, CA - December 11, 2002 ... CoHort Software announced the release of CoStat 6.2, statistical software for scientists and engineers who use Windows or Mac OS X. It includes new procedures for:
CoStat 6.2 includes three unique new approximate procedures for subset selection in multiple regression (Replace2, Simons1, and Simons2). Given a large number of X (independent) variables, subset selection procedures try to find the models with fewer X variables which best explain the Y (dependent) variable. Approximate procedures are needed for large problems for which exact procedures (which test All Subsets) would take years or even thousands of years, even on the fastest computers. There are other approximate algorithms (two are implemented in CoStat: Forward and Replace1; several are not: Backward Elimination, Stepwise, Garrote, Ridge, Forward Stagewise, LARS, and Lasso), but they don't perform nearly as well.
A Monte Carlo test done by CoHort Software with medium-sized problems indicates that CoStat's new procedures are much more likely to find the best models and they do so in a reasonable amount of time.
Procedure Number of Failures Time
(out of 1000 data files) (seconds)
------------------------ ------------------------ ---------
Forward 663 78
Replace1 (like SAS MaxR) 322 134
Replace2 81 292
Simons1 9 578
Simons2 1 1923
All Subsets 0 by definition 43451
The Replace1 procedure (which is the equivalent of MaxR, the best approximate procedure available in SAS), failed with 322 out of the 1,000 data files. In contrast, the new Simons2 procedure, the best approximate procedure available in CoStat, failed with only 1 out of the 1,000 data files.
CoStat has a wide variety of other statistical procedures, too: descriptive statistics, GLM ANOVA, regression (including nonlinear regression), analysis of frequency data, non-parametric tests, etc. The GLM ANOVA procedure can handle missing values, unbalanced designs, and user-specified designs. CoStat can import data from the clipboard, spreadsheet files, ASCII files, and many other types of data files.
CoStat includes a command line version of the program which can process ASCII text commands to run the statistical procedures; or, it can be programmed directly with Java. This makes it ideal for use as a statistics engine on a web server, in batch files, in shell scripts, with pipes, and from Java, Perl, Python and other programs.
CoStat 6.2 is available from CoHort Software in three forms:
CoHort Software of Monterey, California has produced software for scientists and engineers since 1986. For more information, contact CoHort Software (800 728-9878, info@cohort.com, http://www.cohort.com).
JPG screen shots of CoStat are available in two resolutions:
Evaluation copies are available on request.
