Analytics and Data Summit 2019

Finding unexpected patterns and results in data is a critical factor in discovering fraud, adverse reactions, and other suspicious behaviors. Graph analysis is an effective data analysis methodology that considers fine-grained relationships among data entities. This session explains how to use graph analysis algorithms to discover anomalous results to identify malicious network packets in network traces. We achieve this by constructing graphs from network traces and computing characteristics of those graphs by applying several graph algorithms with which we can construct a machine learning model for classification.