Exploring Semantic vs. Syntactic Features for Unsupervised Learning on Application Log Files
Egil Karlsen, Rafael Copstein, Xiao Luo, Jeff Schwartzentruber, Bradley Niblett, Andrew Johnston, Malcolm Iain Heywood, Nur Zincir-Heywood · 2023
This paper explores the effect of traditional TF-IDF and state-of-the-art Large Language models towards anomaly detection in application log files. Four unsupervised learning algorithms are used on three application log files using these models. Results show that each model has advantages and disadvantages depending on the log file used.