МЕТОДИ ТА ЗАСОБИ ВИЗНАЧЕННЯ ОБ'ЄКТІВ РАДІОТЕХНІЧНОЇ РОЗВІДКИ З ВИКОРИСТАННЯМ ТЕХНОЛОГІЙ МАШИННОГО НАВЧАННЯ ТА ОНТОЛОГІЧНОГО ПІДХОДУ

Abstract

The article is devoted to the study of methods and means of determining objects of radio technical intelligence using machine learning technologies and an ontological approach. A naіve Bayesian classifier was used to identify objects of radio technical intelligence. The Naive Bayes classifier is a machine learning algorithm used to classify objects based on probabilities. In this article, a naive Bayesian classifier is used to determine the classes to which objects of radio technical intelligence belong. The classifier uses historical data on object properties to determine the probability that each object belongs to a certain class. For example, based on the properties of the operating frequency range, pulse duration, pulse repetition period, and the number of sources of radio emissions, it is possible to determine the probability that the object belongs to a certain class of radio-technical intelligence. An ontological approach was used to specify the classes to which the objects of radio technical intelligence belong. The ontological approach is used to define classes of objects of radio technical intelligence in order to create a clear and unambiguous model of the subject area. This allows you to structure knowledge about objects, their properties, and relationships, which simplifies further data analysis and allows more accurate classification of new objects. The process of classifying objects in the military field, namely radio-technical intelligence, has been improved by combining the methods of k-nearest neighbors, the naive Bayesian classifier, and the ontological approach, which, unlike the existing methods, before applying the classifier, clustering of objects is carried out in order to take into account the ranges within which features of objects are defined. The analysis of input features showed that the main features for determining the means of radio technical intelligence are: “range of working frequencies”; “impulse duration”; “pulse repetition period”; “the number of sources of radio emissions”. An information system for the classification of radio-technical intelligence tools has been developed, the central component of which is the ontology of radio-technical intelligence tools. Simulation modeling of the work of the developed methods and algorithms was carried out. The choice of software tools for the implementation of the developed methods with the aim of further implementation on various platforms is substantiated. The JavaScript programming language using the JQuery library was used to implement the functional content of the system. The conducted simulation shows a satisfactory result of the developed methods and algorithms.