Aim

This was my final year project for University, for this we developed an application that was able to take log files and determine the root cause of the issue.

Abstract

"In this paper, we perform anomaly detection and root cause analysis on log data (specifically system logs). Firstly, we employ a log parsing solution known as Drain (an online log parsing approach with fixed depth tree). We then present an anomaly detection approach that utilizes a decision tree model. This will be used to determine the anomalous devices in the log files. One benefit of decision tree models is that they are easily traceable, providing a contrast to most “black-box” solutions currently available in the industry. Finally, a sequential model using Keras is built to predict the root cause of a given issue."

Outcome

Our paper is published in Artificial Intelligence XXXIX