Network Anomaly Detection with Compression

Anomaly detection suffers from the dynamic web data with more false alarms or frequent necessity of model retraining. In this paper, we propose a framework which makes a multidisciplinary cooperation between compression and variational segment model. The detection model employ compression based probability estimation and fine-grained structure analysis of the web request. We compare our model with state-of-art detectors using real-world dataset. The experimental results prove the performance of our model with high detection accuracy and low false alarms.