Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back Increase of convergence of training processes in diagnostic models of systems based on procedures of ranking V.I. SENCHENKOV, I.N. NEKRASOV Evolve the theoretical bases diagnosis of complex technical systems on the basis of training methods. From the new positions the topology of the set of observed states and related issues of image construction are revealed - a formal representation of the types of technical condition of the system. A procedure for ranking training images is developed, which makes it possible to significantly accelerate the convergence of the training process. This procedure is based on the analysis of metric relationships in clusters of training images and the assignment on this basis of the order of their application in the training process. The metric of finite-dimensional Euclidean spaces is used. An example of the ranking of training images based on the proposed procedure is given. Comparison of the rate of convergence of training processes with the application of this procedure and approaches to the processing of statistical information is made, which were proposed in previous publications. Keywords: technical condition, attitude of tolerance, training process, cluster, ranking procedure. Contacts: E-mail: svi9@rambler.ru Pp. 14-21.