Performance Comparison of Malware Classification Algorithms using WEKA Tool

Monitoring a computer network and protecting the resources within it is a topic of great interest for a wide range of applications such as those used in the military, finance, healthcare, and many others. Also, in addition to effective monitoring, a continuous development of cyber security tools is also necessary because attacks are becoming more frequent and attackers are also developing more and more mechanisms to steal information and affect services which, at first glance, would be of extremely low interest to anyone. Within the work, a classification of pre-tagged network traffic will be performed to detect malware applications. The experiments will comprise two distinct approaches: a binary classification approach, to simply distinguish between normal and malicious traffic, and a multi-class classification approach, to detect different types of malicious traffic. The way of testing these two types of databases will be done using the WEKA tool with two ways of splitting the data - percentage split type $80 \%-20 \%$ and type $70 \%-30 \%$. Afterwards, a comparison will be realized, in terms of efficiency and accuracy, between the Machine Learning algorithms used in the experiments.