Fraudulent IP Address Detection Using Machine Learning Techniques

The popularity of network technology and the threats involved make it imperative to develop various techniques to effectively detect fraudulent activities. The focus of this paper is to discover actionable irregularity with significant potential to perceive abnormal network behavior. Three machine learning-based methods have been proposed for the analysis of suspicious network behavior. These methods are extensions of the techniques discussed as part of the introduction below. The method developed in the current work can be evaluated by testing its efficiency against real-time network attacks using available open-source network tools. Experimental results show that irregularities have been successfully identified from the dataset with a low false positive rate. Furthermore, we believe our method can be directly deployed in real-time environments (either independently of edge devices or via the cloud) to strengthen network security.