Open Set Recognition with Dissimilarity Weight for Unknown Attack Detection

As information technology advances, it provides user convenience but also has more vulnerabilities than ever before. In particular, attackers use advanced techniques to perform new attacks. cyber security, such attacks are defined as unknown and target previously undetected or excavate gaps in the system. Because these unidentified unanalyzed, they difficult identify signature-based misuse detection that learns rules patterns. Furthermore, anomaly-based from normal data detect outliers cannot accurately, because does not distinguish between known To overcome problems, this study applied Open-Set Recognition with dissimilarity weight(OSRDW). A OSRDW method was used effectively train extreme value distribution, which calculated by applying weight, through attack’s weights were classified Through research analysis, attack divide two types, three sets(NSL-KDD, UNSW-NB15, CICIDS-2017) experiment. For first type of attack, rate proposed approximately 10%–20% better conventional method. second accuracy higher for The experimental results confirmed had performance detecting detected various sets.