Improved N-gram approach for cross-site scripting detection in Online Social Network

Nowadays Online Social Networks (OSNs) have become a popular web service in the world. With the development of mobile networks, OSNs provide users with online communication platform. However, the OSNs´ openness leads to so much exposure that it brings many new security threats, such as cross-site scripting (XSS) attacks. In this paper, we present a novel approach using classifiers and the improved n-gram model to do the XSS detection in OSN. Firstly, we identify a group of features from webpages and use them to generate classifiers for XSS detection. Secondly, we present an improved n-gram model (a model derived from n-gram model) built from the features to classify webpages. Thirdly, we propose an approach based on the combination of classifiers and the improved n-gram model to detect XSS in OSN. Finally, a method is proposed to simulate XSS worm spread in OSN to get more accurate experiment data. Our experiment results demonstrate that our approach is effective in OSN´s XSS detection.