MATLAB Implementation of the Fukunaga Koontz Transform

MATLAB Implementation of a feature selection and ordering algorithm developed by Fukunaga and Koontz

The Fukunaga Koontz transform (FKT) is a linear transform that can be applied to a
dataset X derived from a set of objects that are each classified into one of two classes. X is
an N x L matrix where each row corresponds to an object and each column is a real valued
attribute of the object. In the transformed space the attributes that are the most important for class 1 are the least for class 2 and vice versa.