Засіб для біометричної автентифікації на основі поведінкових особливостей користувача

Косарева, А. С.; Регіда, П. Г.

dc.contributor.author	Косарева, А. С.
dc.contributor.author	Регіда, П. Г.
dc.date.accessioned	2021-10-27T15:55:51Z
dc.date.available	2021-10-27T15:55:51Z
dc.date.issued	2021
dc.identifier.uri	http://ir.stu.cn.ua/123456789/24134
dc.description	Косарева, А. С. Засіб для біометричної автентифікації на основі поведінкових особливостей користувача / А. С. Косарева, П. Г. Регіда // Технічні науки та технології. – 2021. – № 2(24). – С. 114-122.	en_US
dc.description.abstract	У роботі наведено результати дослідження існуючих методів поведінкової біометрії, а саме біометрії за особливостями користувача при наборі тексту. Також у роботі описано та проаналізовано власний метод автентифікації користувача за його динамікою натискання клавіш. Запропоновано розроблений авторами метод, описаний у статті, як програмний додаток для проведення різноманітних досліджень розпізнавання особистості за її поведінковими особливостями.	en_US
dc.language.iso	uk	en_US
dc.publisher	Чернігів : НУ «Чернігівська політехніка»	en_US
dc.relation.ispartofseries	Технічні науки та технології;№2(24)
dc.subject	біометрія	en_US
dc.subject	автентифікація	en_US
dc.subject	поведінкові особливості	en_US
dc.subject	динаміка натискання клавіш	en_US
dc.subject	нейронні мережі	en_US
dc.subject	biometrics	en_US
dc.subject	authentication	en_US
dc.subject	behavioral characteristics	en_US
dc.subject	keystroke dynamics	en_US
dc.subject	neural networks	en_US
dc.title	Засіб для біометричної автентифікації на основі поведінкових особливостей користувача	en_US
dc.title.alternative	Tool for biometric authentication based on user behavioral features	en_US
dc.type	Article	en_US
dc.description.abstractalt1	Biometric authentication is one of the most common ways to identify a person. However, the topic of behavioral biometrics is almost unstudied, and those authentication methods can significantly increase the level of security of personal data, but it is still not implemented in modern devices due to lack of an existing researches and scientific works. It is a very important issue to ensure the integrity of personal data by hackers or attackers, so password recognition is being supplemented by biometric authentication. Today, the most common methods are to recognize users by fingerprint or facial geometry. But even such personal characteristics can be made public. Therefore, research on behavioral biometrics is essential, as such authentication methods do not require additional effort from the user and can complement the security system. There are currently such behavioral biometrics methods that use keyboard typing dynamics, keystroke type strength, typing speed, and other criteria studied in described papers by authors such as Abdulaziz Ali Alzubaidi, Yugal Kalita, Troyan and Ortmeyer, and others. Since the methods of user recognition by its behavioral characteristics are not studied enough in the field of biometric authentication, one of the goals is to create authors’ own method of user recognition, analysis and comparison with existing solutions. The aim of the paper is to create a platform for analyzing various methods of behavior authentication and its spreading through open source policy. This will help researchers and scientists to conduct their own experiments and improve the level of protection by behavioral biometrics methods. The created method of user recognizing by their behavioral features is described, namely by means of dynamics of keystrokes, speed of typing, angle of inclination of a smartphone during its use. The neural network, which is trained for personality recognition, responds to the generated data packet to run the function, the results of which are recorded and displayed in the form of graphs of the probabilities of the Accept response and the Reject response. The system of user recognition by their behavioral features, which is provided in open source, is developed, the created program is analyzed and results of its work in the form of graphs are provided. It was concluded that the maximum allowable error rate is 2000 data occurrences.	en_US