*********************************************************************************
* Generalized Disciminant Analysis for MataLab version for Windows 5.2.0.3084 	*
*										*
* Gaston Baudat & Fatiha Anouar / 21st October 2000 / Exton PA 19341 USA	*
*********************************************************************************

This MatLab (5) version of the GDA uses the method of resolution presented in:

"Generalized Discriminant Analysis Using a Kernel Approach"  	(Baudat, Anouar 2000)
Neural Computation  12,2385-2404 (2000)

Need the following files (*.m):

	KernelFunction.m	The kernel current function.
	EigenSystem.m		Evaluate and sort the eigen values and vectors.
	DataSt.m		Center and normalize the data.
	Iris.m			The Fisher's iris data (3 x 50 samples / 4 variables)
	BuildGDA.m		Build the GDA solution (give a data structure)
	SpreadGDA.m		Spread test vectors into the GDA discriminant subspace.
	PlotGDA.m		Plot data into one, or two, discriminant axes.

	GDA_Iiris.m		An example of the GDA with the Fishier's iris data

The GDA_Iris.m file gives an example of the GDA using the classic Fishier's iris data.
To test the user can compare the results with (Baudat, Anouar 2000).
The orignial kernel was a Gaussian with sigma = 0.7.

Gaston Baudat & Fatiha Anouar




Any questions, or comments can be sent to:

gbaudat@voicenet.com or fanouar@voicenet.com