Publication: Prediction of intrusion detection in voice over internet protocol system using machine learning
Issued Date
2021-01-01
Resource Type
Other identifier(s)
2-s2.0-85121292237
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
ACM International Conference Proceeding Series. (2021), 149-155
Suggested Citation
Chaiyos Choti, Narit Hnoohom, Suratose Tritilanunt, Sumeth Yuenyong Prediction of intrusion detection in voice over internet protocol system using machine learning. ACM International Conference Proceeding Series. (2021), 149-155. doi:10.1145/3479162.3479185 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/76728
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Prediction of intrusion detection in voice over internet protocol system using machine learning
Other Contributor(s)
Abstract
Currently, the threats of the internet and computer network have still at a high level. The firewall installation does not enough to detect intrusion of any attempts into the network. This research developed an intrusion detection that corresponded to the attack behavior by inspecting the components of SIP messages from the network traffic. The rules determination applied to detect intrusion related to unauthorized access, sending fraudulent SIP messages for denial of service, and flooding with malformed packets to reduce the performance of the network operation. The data collection would be from actual operation of the system in scenarios of attack and normal activity. The performance measure of data classification using machine learning algorithms shows that the support vector machine had the lowest accuracy of 90.81%, artificial neural network had an accuracy of 96.92%, decision tree had an accuracy of 99.29%, k-nearest neighbor had the highest accuracy of 99.81%. The evaluation of time consumed for training a dataset and detection per a data item shows that the decision tree spent the least time of 0.014s and 0.000ms, support vector machine spent 1.276s and 0.000ms, k-nearest neighbor spent 2.535s and 0.001ms, whereas the artificial neural network spent the most time of 7.425s, and 0.005ms, respectively.