5G Networks Security: Attack Detection Using the J48 and the Random Forest Tree Classifiers
Steele II, Bruce
Kholidy, Hisham A.; Advisor
SUNY Polytechnic Institute
5G networks , radios , next generation , millimeter waves , network function virtualization , network slicing , cellular communication networks
5G is the next generation of cellular networks succeeding and improving upon the last generation of 4G Long Term Evolution (LTE) networks. With the introduction of 5G comes significant improvements over the previous generation with the ability to support new and emerging technologies in addition to the growth in the number of devices. The purpose of this report is to give a broad overview of what 5G encompasses including the architecture, underlying technology, advanced features, use cases/applications, and security, and to evaluate the security of this new networks using existing machine learning classification techniques such as The J48 Tree Classifier and the Random Forest tree classifier. The evaluation is based on the UNSW-NB15 dataset that was created at the Cyber Range Lab of the Australian Centre for Cyber Security (ACCS) at the University of New South Wales. Since 5G datasets have yet to have been created, there is no publicly available dataset for the 5G systems. However, While the UNSW-NB15 dataset is built using a standard wireless computer network, we will use it to simulate the device-to-device (D2D) connections that 5G will support. In the case with the UNSW dataset, the J48 tree classifier fits more accurately than the Random Forest classifier. The J48 tree classifier achieved an 86.422% of correctly classified instances. On the other hand, the Random Forest tree classifier achieved 85.8451% of correctly classified instances.