Engineering Physics Annotation << Back NEW TRAFFIC CLASSIFICATION METHOD USING HIDDEN MARKOV MODEL A.M. ANDREEF, A.V. BABICHENKO, S.M. JAMMOUL Network traffi c classifi cation is one of the most highly researched topics in last decade. Most of the proposed methods do not rely on clear mathematical models. In work presents new traffi c classifi cation method using hidden Markov model and new enhancements in choosing model’s initial values. Tests the presented method has been done using four applications (HTTP, HTTPS, IMAPS and WhatsApp) as well a real traffi c logs from BMSTU – Bauman University. The method shows a very good results, it identify the application in real time with precision > 95 %. Key words: network traffic models; traffic classifi cation; traffic identification; hidden markov model; internet application; markov models. DOI: 10.25791/infizik.06.2018.047 Contacts: E-mail: arkandreev@gmail.com, E-mail: ababichenko@rpkb.ru Pp. 38-47.