You are here:
Quick Overview Scout 2008 Version 1.0
Citation:
Singh, A. AND J. M. NOCERINO. Quick Overview Scout 2008 Version 1.0. US Environmental Protection Agency, Cincinnati, OH, EPA/600/F-08/016, 2008.
Impact/Purpose:
Several embedded commercial software programs were evaluated to shorten development time and to ensure the longevity of Super 2008. Scout 2008 has been developed in the Microsoft .NET Framework using the C# programming language to run under the Microsoft Windows XP operating systems. As such, to properly run Scout 2008, the computer using the program must have the .NET Framework pre-installed. Most computers running under Windows XP has this framework installed; however, if there are problems encountered when trying to run Scout 2008, then check to make sure that NET Framework version 1.1 (or later) is installed.
Description:
The Scout 2008 version 1.0 statistical software package has been updated from past DOS and Windows versions to provide classical and robust univariate and multivariate graphical and statistical methods that are not typically available in commercial or freeware statistical software packages. Scout 2008 version 1.0 runs under the Microsoft Windows XP operating system (see the section on “Software used to develop Scout 2008” for details). Scout 2008 includes the latest state-of-the science classical, robust, and resistant univariate and multivariate outlier identification and robust estimation methods available, including: iterative classical, iterative influence function (e.g., biweight, Huber, PROP) based M-estimates, multivariate trimming (MVT), least medianof- squared residuals (LMS) regression, and minimum covariance determinant (MCD). Scout 2008 offers classical and robust methods to estimate: multivariate location and scale, univariate intervals, multiple linear regression parameters, principal components (PCs), and discriminant (Fisher, linear, and quadratic) functions (DFs). The discriminant analysis module of Scout 2008 can perform cross validation using several methods, including: leave-one-out (LOO), split samples, and bootstrap methods. Some initial choices for the iterative estimation of location and scale include: the orthogonalized Kettenring and Gnanadesikan (OKG) method; median, median absolute deviation (MAD), or inter-quartile range (IQR) based estimates; and the MCD method.