An unsupervised Bayes classifier for normal patterns based on marginal densities analysis

摘要

In this paper, an approach to unsupervised pattern classification is discussed. The classification scheme is based on an approximation of the probability densities of each class under the assumption that the input patterns are of a normal mixture.The description of the marginal densities in terms of convexity allows one to determine, from a totally unlabelled set of samples, the number of components and, for each of them, approximate values of the mean vector, the covariance matrix and the a priori probability. Discriminant functions can then be constructed. Computer simulations show that the procedure yields decision rules whose performance remains close to the optimum Bayes minimum error-rate, while involving only a small amount of computation.