作者: Liu Bo , Wang Zhanyi , Zhang Zhuo
DOI:
关键词: Computer science 、 Parsing 、 Anomaly detection 、 Anomaly (physics) 、 Feature (computer vision) 、 Data mining 、 Detection rate 、 Web access
摘要: A web anomaly detection method and apparatus. The comprises: according to a plurality of access features parsed from historical record, creating models for detecting an anomalous based on the features; adopting various respectively detect whether target is access; labelling feature corresponding model in which result as type access. In embodiments present invention, created are utilized automatically unknown loopholes new attacks, without re-collecting data making rules, thereby guaranteeing relatively high rate accuracy rate, reducing manual workload saving manpower time.