Non parametric population classification

Population classification is a set of statistical techniques to classify populations on basis of observations of their constituing members. This decision making involves a cascade of classifiers: one for objects, and, based on the resulting classifications, one for populations. Also a rule is to be given how many objects should be analysed before the population is decidable. In search for a more integral view of the problem the population function is introduced. Consideration of the population function implies that rather then fixing the object classifier a priori, it is more efficient and more accurate to extend the classifier to a distinct range of the feature vector. The concept of sequential classification is defined as well. Population classification performance is favourably compared to [1,2,3,5,6].