Next: Results and discussion
Up: Methods
Previous: Parzen windows
Fisher's linear discriminant is a classification method that projects
highdimensional data onto a line and performs classification in this
onedimensional space. The projection maximizes the distance between
the means of the two classes while minimizing the variance within each
class. This defines the Fisher criterion, which is maximized over all
linear projections, w:

(9) 
where m represents a mean, s^{2} represents a variance, and the
subscripts denote the two classes. In signal theory, this criterion is
also known as the signaltointerference ratio. Maximizing this
criterion yields a closed form solution that involves the inverse of a
covariancelike matrix. This method has strong parallels to linear
perceptrons. We learn the threshold by optimizing a cost function on
the training set.
Michael Brown
19991105