ONE-CLASS FUSION-BASED LEARNING MODEL FOR ANOMALY DETECTION

The Dempster-Shafer (DS) theory of evidence is frequently used to combine multipe supervised machine learning models into a robust fusion-based model. However, using the DS create fusion model from multiple one-class classifications (OCCs) for network anomaly detection challenging task. First, lack attack data leads difficulty in estimating an appropriate threshold OCC distinguish between normal and abnormal samples. Second, it also very find weight OCCs that corresponds contribution each In this paper, we attempt solve above issues order make applicable constructing OCC-based models. Specifically, propose two novel methods automatically choosing individual Thanks that, develop One-class Fusion-based Anomaly Detection (OFuseAD) single OCCs. proposed evaluated on ten well-known problems. experimental results show performance OFuseAD improved almost all tested datasets metrics: accuray F1-score. visualization provides insight characteristics OFuseAD.