Open AccessMethodical approach to assess the condition of samples of weapons and military equipment on the basis of algorithmic trees
Author(s) -
V. Dudnyk,
O. Grishchyn,
V. Netrebko,
R. Prus,
M. Voloshcuk
Publication year - 2021
Publication title -
vìjsʹkovo-tehnìčnij zbìrnik
Language(s) - English
Resource type - Journals
eISSN - 2708-5228
pISSN - 2312-4458
DOI - 10.33577/2312-4458.25.2021.69-76
Subject(s) - set (abstract data type) , basis (linear algebra) , tree (set theory) , computer science , one class classification , sample (material) , data mining , artificial intelligence , process (computing) , decision tree , machine learning , statistical classification , task (project management) , algorithm , pattern recognition (psychology) , mathematics , support vector machine , engineering , systems engineering , mathematical analysis , chemistry , geometry , chromatography , programming language , operating system
An effective mechanism for the synthesis of classification trees based on fixed initial information (in the form of a training sample) for the task of recognizing the technical condition of samples of weapons and military equipment. The constructed algorithmic classification tree (model) will unmistakably classify (recognize) the entire training sample (situational objects) according to which the classification scheme is constructed. And have a minimal structure (structural complexity) and consist of components (modules) - autonomous algorithms for classification and recognition as vertices of the structure (attributes of the tree). The developed method of building models of algorithm trees (classification schemes) allows you to work with training samples of a large amount of different types of information (discrete type). Provides high accuracy, speed and economy of hardware resources in the process of generating the final classification scheme, build classification trees (models) with a predetermined accuracy. The approach of synthesis of new algorithms of recognition (classification) on the basis of library (set) of already known algorithms (schemes) and methods is offered. Based on the proposed concept of algorithmic classification trees, a set of models was built, which provided effective classification and prediction of the technical condition of samples. The paper proposes a set of general indicators (parameters), which allows to effectively present the general characteristics of the classification tree model, it is possible to use it to select the most optimal tree of algorithms from a set based on methods of random classification trees. Practical tests have confirmed the efficiency of mathematical software and models of algorithm trees.