期刊名称：International Journal of Computer and Information Technology
印刷版ISSN：2279-0764
出版年度：2017
卷号：6
期号：6
页码：277-285
出版社：International Journal of Computer and Information Technology
摘要：Advancements of technology in the field of networking
and computing bring with it a heightened importance for cyber
security. Currently, legal forms of identification, financial
information, and scientific data rely on this technology.
Moreover, as more dependence on cloud computing, data-base
storage, and online banking is applied keeping sensitive
information secure is paramount. JavaScript Cross-site based
attacks continue to be the most prevalent cause of compromised
information. Here, we demonstrate the feasibility of using
unsupervised learning algorithms to detect attacks as part of the
Intrusion Detection System for malicious cross scripts with
attacked web sites. Our contribution is domain-based, in order to
track changes of the interaction. Profiles are made from webcrawled
pages and parsed according to key scripting features.
The detection is done when scripting deviates from the main
clustering of those clean profiles data gathered.
