Real-Time Threat Detection in UEBA using Unsupervised Learning Algorithms

In this modern world of digital communications and transactions, cybersecurity and the protection of user data have been of utmost importance. User and Entity behavior analytics is a powerful tool to prevent various threats. Through this paper, we bring to you a proposed machine learning UEBA model which protects user data and prevents insider threats more efficiently. We have tried to compare four different unsupervised algorithms which we believe to be far superior to the normally supervised machine learning algorithms. Our main aim is to provide a more efficient UEBA model through the combination of the above-mentioned algorithms. On comparing with the normally used supervised algorithms, we have observed that our proposed model works much more efficiently and is less time-consuming.