Open AccessMETHOD OF SPAM DETECTION BASED ON ARTIFICIAL IMMUNE SYSTEMS
Author(s) -
M. P. Malykhina,
В А Частикова,
Alexandr Aleksandrovich Biktimirov
Publication year - 2018
Publication title -
vestnik astrahanskogo gosudarstvennogo tehničeskogo universiteta. seriâ: upravlenie, vyčislitelʹnaâ tehnika i informatika
Language(s) - English
Resource type - Journals
eISSN - 2224-9761
pISSN - 2072-9502
DOI - 10.24143/2072-9502-2018-3-38-48
Subject(s) - artificial immune system , computer science , software , heuristic , artificial intelligence , task (project management) , machine learning , systems engineering , engineering , programming language
The task of developing tools to combat spam is currently focused on creating such techniques for detecting spam, which are endowed with the skills and qualities inherent in a person whose work is not limited to patterns and therefore highly effective. Man has the ability to detect spam signs, which is based on his own knowledge, experience and preferences. There has been substantiated the need to develop a new approach to solving the problem of detecting spam messages, which is based on heuristic methods of optimization, is effective at the initial stage of training and has a low frequency of false operations. This formulation of the problem fully corresponds to modeling mechanisms of the immune systems of living organisms that ensure their survival, these mechanisms being represented, investigated and used by software. There have been identified and described main mechanisms of artificial immune systems intended for solving the problem of spam detection, as well as software and system interacting. The basic concepts of constructing an artificial immune system for the purpose formulated above are determined: class of detectors, presentation of receptors and pathogens. A model of the relationships between them has been worked out. A technique for detecting spam based on the work of an artificial immune system is proposed, an algorithm for its implementation is developed, and the specifics of its members to identify spam messages are described. A software package with advanced research capabilities has been created. Testing and analysis of the results to determine the optimum values of the system operation parameters have been conducted.