ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 10. Vol. 25. 2019

DOI: 10.17587/it.25.602-608

Yu. A. Mezentsev, Doctor of Technical Sciences, Professor, e-mail: mesyan@yandex.ru,
O. M. Razumnikova, Doctor of Biological Sciences, Professor, e-mail: razoum@mail.ru, Novosibirsk State Technical University, Novosibirsk, Russia,
I. V. Tarasova, Doctor of Medical Sciences, Leading Researcher, e-mail: iriz78@mail.ru,
O. A. Trubnikova, Doctor of Medical Sciences, Head of Laboratory, e-mail: olgalet17@mail.ru, Research Institute for Complex Issues of Cardiovascular Diseases, Kemerovo, Russia

On some Problems of Big Data Clustering by Minimax and Additive Criteria, Application in Medicine and Neurophysiology

The NP-hard clustering problem as applied to the data of neurophysiological studies (indicators of postoperative cognitive dysfunction) is considered. Variants of the problem of clustering in the form of mixed integer programming, including the use of continuous relaxation, reducing the complexity of solutions without loss of accuracy are given. The results of computational experiments on real data using the software implementation of the algorithm of binary cuts and branch are presented. They demonstrate the high efficiency of the developed toolkit.
Keywords: clustering, minimax quality criterion, additive criterion, linear relaxation, binary cutoff and branching algorithm, detection of postoperative cognitive dysfunction

P. 602–608

To the contents