Denial of Service Attack Prediction Using Gradient Descent Algorithm

Gayathri Rajakumaran, Neelanarayanan Venkataraman, Raghava Rao Mukkamala

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review


Denial of service (DoS) attack is one of the prevalent security threats in today’s digital world. A significant number of machine learning algorithms have been applied for detection of DoS attacks. However, each algorithm has its own limitations. In general, the success of any machine learning algorithm is based on the selection of appropriate data set and identification of attack parameters. In this paper, a detailed investigation is done in the process of identifying relevant attack parameters from the simple network management protocol data set. The chosen parameters underwent various metrics comparisons for validating their accuracy. We started with the linear regression model and achieved accuracy of 99.7% with 3.3% errors. Hence, to achieve further optimization in the case of error reduction, we applied gradient descent algorithm in the linear regression which reduces errors by 3%. Hence, our proposed measures help in accurate identification of DoS attacks and the same has been verified through the experimental simulations and graphical representation.
TidsskriftSN Computer Science
Udgave nummer1
Antal sider8
StatusUdgivet - jan. 2020


  • DoS
  • SNMP
  • Gradient descent