作者: Ming-Tian Zhou , Zhi-Hong Zuo , Ke Tang
DOI:
关键词: Similarity (geometry) 、 False positive rate 、 Metric (mathematics) 、 Signature (logic) 、 Computer science 、 Centroid 、 Algorithm 、 Mahalanobis distance 、 Malware 、 Cluster analysis
摘要: Polymorphic malware is a secure menace for application of computer network systems because hacker can evade detection and launch stealthy attacks. In this paper, novel enhanced automated signature generation (EASG) algorithm to detect polymorphic proposed. The EASG composed enhanced-expectation maximum K-means clustering algorithm. algorithm, the fixed threshold value replaced by decision interval area. false positive ratio be controlled at low level, iterative operations execution time are effectively reduced. Moreover, centroid updating realized similarity metric Mahalanobis distance incremental learning. Different group families partitioned updating.