On Solving Problems of Classification of Graphic Images in the Wolfram Mathematica System

Abstract

In recent years, problems solved through intelligent technologies have become widespread. Therefore, modern computer mathematics systems also include tools that allow solving problems through the use of artificial intelligence technologies. In particular, this includes solving problems of classification of graphic images, that is, assigning a graphic image to a particular class. When solving problems of this type, machine learning is used. Tools for solving classification problems, in particular, graphic images, are included in the latest versions of the Wolfram Mathematica system. The Wolfram Mathematica system makes it possible to build a classifier function, and using various algorithms, to determine the probabilities of an image belonging to a particular class, to build a classification error matrix, to output other information about the properties of the classifier. In addition to graphical images, classifiers can be built for numeric, text, and other data types. The paper considers the applications of these new capabilities of the Wolfram Mathematica system to solving problems that arise in the practice of air defense: recognition of unmanned aerial vehicles by their images and recognition of air targets by their radar range portraits. These tasks are of great applied importance, and the proposed method of solving them using Wolfram Mathematica can be used both in the course of scientific research and when familiarizing military school cadets with the capabilities of modern artificial intelligence technologies. At the same time, the use of the Wolfram Mathematica system allows you to use its standard interface, which students or researchers previously got acquainted with when solving other problems.

Keywords

• classification of graphic images
• wolfram mathematica system