On an optimal linear pattern classification procedure
暂无分享,去创建一个
Abstract The problem of Optimal Linear Discrimination for 3-class pattern recognition is attacked. The Linear Discrimination Rule, under those circumstances, is devised according to the following optimality criterion: maximize the probability of correct classification into any one of the three classes, while keeping the other two such probabilities fixed. An explicit solution of the problem is given, and a search technique is implemented greatly reducing the problem complexity.
[1] T. W. Anderson,et al. Classification into two Multivariate Normal Distributions with Different Covariance Matrices , 1962 .
[2] D. W. Peterson,et al. A method of finding linear discriminant functions for a class of performance criteria , 1966, IEEE Trans. Inf. Theory.
[3] L. Mirsky,et al. The Theory of Matrices , 1961, The Mathematical Gazette.