Study of the features of a multi-element classifier for unobservable classes based on a gaussian mixture distribution model
2026, vol.18 , no.4, pp. 59-66
Article [2026-04-06]
The article is devoted to the study of the features of a multi-element classifier based on a model of a Gaussian mixture of distributions. The relevance is determined by the widespread development of intelligent data processing systems based on machine learning technologies. The choice of a classifier model based on a Gaussian mixture of distributions is due to its ability to represent a large class of sample distributions. A classifier model based on a Gaussian mixture of distributions is proposed to solve the problem of multiclass image classification. The proposed model is two-level: the level of image preprocessing; the level of multi-element image classification. An algorithm for the functioning of a classifier of a multi-element classification system operating under conditions of incompleteness of classes of local classifiers has been developed.
the features, multi-element classifier, Gaussian mixture distribution model, machine learning, Federative machine learning, unobservable classes
https://doi.org/10.59035/WPIQ2651
O. Ja. Kravets, O. Yu. Zaslavskaya, P. V. Bespalov, Yu. V. Redkin, O. O. Varlamov. Study of the features of a multi-element classifier for unobservable classes based on a gaussian mixture distribution model. International Journal on Information Technologies and Security, vol.18 , no.4, 2026, pp. 59-66. https://doi.org/10.59035/WPIQ2651