Applied Statistics (2nd edition) CD

1  CD Contents

The contents of the CD are as follows:

·  Datasets in the form of Excel files for statistical experiments.

·  Tools for specific statistical procedures.

Both datasets and tools are described in detail in the book.

2  System Requirements

·  PC with CD player, running Windows 98 SE, Windows NT 3 SP4 or later with Microsoft Excel 2003.

·  A minimum of 17 Mbytes should be available in the Hard Disk in order to install all datasets (14.1 Mbytes) and tools (2.4 Mbytes).

·  A minimum of 128 Mbytes RAM and 550 MHz processor rate are recommended for acceptable performance.

·  Matlab

3  Installation

·  Datasets and Tools.xls are simply copied from the CD to the destination directory.

·  Matlab functions are usually copied from the CD to the Matlab "work" directory. Other alternative directories are also possible (see Matlab manual).

·  R functions are text files that should be pasted into the R console.

·  The SCSize tool is provided with a setup.exe file for installation by either running the setup.exe file from the Run option of the Start menu or by using the Add/Remove Programs option of the Control Panel option in Settings. Next, simply follow the instructions. Before installation it is advisable to first close all open programs.

4  Using SCSize

SC Size (Statistical Classifier Design)

Displays a graphic of the following variables, for a two-class linear classifier with specified Battacharrya distance and for several values of the dimensionality ratio (number of patterns per class/dimension):

·  Bayes error;

·  Expected design set error (resubstitution method);

·  Expected test set error (holdout method).

The standard deviations of the error estimates are also displayed when the mouse is clicked over a selected point of the picture box.

The user must specify:

·  Dimension d (£ 10).

·  Square of the Battacharrya distance, computable by several statistical software products, and usually called square Mahalanobis distance (of the class means).