Discriminant analysis of binary data following multivariate Bernoulli distribution

A new method of discriminant analysis of classifying binary data is proposed by considering an exact joint probability mass function of correlated binary variables. The interaction order of the joint probability mass function can be controlled for the performance of the proposed method on the classification accuracy and computational time. The performance in terms of the misclassification rate for a real data and some artificial data sets was reported and compared with those of linear discriminant analysis and logistic regression.