Using multidimensional pattern recognition techniques in classifying primary cosmic-ray particles

The application of statistical multidimensional pattern recognition theory to solving problems associated with the classification of primary cosmic-ray particles is studied. As an illustration, a practical problem that has not even an approximate solution within the deterministic approach is solved on the basis of model data. An algorithm that is used to divide the primary particles into two classes and employs the concept of a Bayes linear classifier is described, and a broad spectrum of problems of experimental space physics is specified whose solution can be achieved by using the proposed method.