ABSTRACT
Cloud computing is a cost-effective way to host and deliver services over the internet, but the large volume of data transmitted makes the cloud network a target for malicious attacks. To protect against these threats, various intrusion detection systems (IDSs) have been developed to identify different types of attacks on the network. However, some of these systems are not always efficient due to the level of human involvement they require. Machine learning approaches, which have been widely used in IDSs, have the potential to improve efficiency by building well-trained models for anomaly detection using large datasets with a variety of attack types. In this paper, we examine the current state of research on using machine learning techniques for IDSs and evaluate the suitability of different algorithms using the UNSW dataset. We compare the accuracy rate and classification performance of several algorithms to determine their effectiveness.
