Behavioral detection of spam URL sharing: posting patterns versus click patterns

摘要: Social media systems like Twitter and Facebook provide a global infrastructure for sharing information, in one popular direction, of web hyperlinks. Understanding the behavioral signals both how URLs are inserted into these (via posting by users) received social users clicking) can new insights search, recommendation, user profiling, among many others. Such studies, however, have traditionally been difficult due to proprietary (and sometimes private) nature much URL-related data. Hence, this paper, we begin examination URL through two distinct perspectives: (i) first is via study links posted publicly-accessible data; (ii) second measuring their click patterns Bitly API. We examine differences between sample application domain: classification spam URLs. find that - versus clicking overlapping but fundamentally different perspectives on URLs, inform design future applications link detection sharing.