INTELLIGENT MODEL FOR CLASSIFYING NETWORK CYBERSECURITY EVENTS
Тетяна Василівна Бабенко, Andrii Bigdan, Larisa Myrutenko · Information systems and technologies security · 2023
Due to the increased complexity of modern computer attacks, there is a need for security professionals not only to detect harmful activity but also to determine the appropriate steps that an attacker will go through when performing an attack. Even though the detection of exploits and vulnerabilities is growing every day, the development of protection methods is progressing much more slowly than attack methods. Therefore, this remains an open research problem. In this article, we present our research in network attack identification using neural networks, in particular Rumelhart's multilayer perceptron, to identify and predict future network security events based on previous observations. To ensure the quality of the training process and obtain the desired generalization of the model, 4 million records accumulated over 7 days by the Canadian Cybersecurity Institute were used. Our result shows that neural network models based on a multilayer perceptron can be used after refinement to detect and predict network security events.