A new method for the extraction of features in a two-class pattern recognition problem is derived. The main advantage is that the method for selecting features is based entirely upon discrimination or separability as opposed to the more common approach of fitting. The classical example of fitting is the use of the eigenvectors of the lumped covariance matrix corresponding to the largest eigenvalues. In an analogous manner, the new technique selects discriminant vectors (or features) corresponding to the largest "discrim-values." The new method is compared to some of the more popular alternative techniques via both data-dependent and mathematical examples. In addition, a recursive method for obtaining the discriminant vectors is given.

This paper describes an on-line interactive graphics system which has been designed to solve the problems of pattern analysis and pattern classification. A wide variety of both classical and unique mathematical algorithms, along with their graphic system implementation, are discussed. A discussion of an application of the system to the handprinted character recognition problem is included.

In solving pattern classification problems, many researchers have successfully used the Fisher linear discriminant as the optimal linear method for discriminating between vector samples from two classes. With the introduction of on-line, interactive, graphic systems, it has become conveniently possible to extend the discrimination logic to piecewise linear methods. This paper describes such a method which is being used in the on-line pattern analysis and recognition system (OLPARS).

An algorithm for the analysis of multivariate data is presented along with some experimental results. The algorithm is based upon a point mapping of N L-dimensional vectors from the L-space to a lower-dimensional space such that the inherent data "structure" is approximately preserved.