Description

Ranking fraud in the mobile App market refers to fraudulent or deceptive activities which have a purpose of bumping up the Apps in the popularity list indeed, it becomes more and more frequent for App developers to use shady means, such as inflating their Apps’ sales or posting phony App ratings, to commit ranking fraud. While the importance of preventing ranking fraud has been widely recognized, there is limited understanding and research in this area. Several recent studies have pointed out that advertising in mobile apps is plagued by various types of frauds. In Web advertising, most fraud detection is centered on analyzing server-side logs network traffic, which are mostly effective for detecting bot-driven ads. These can also reveal placement frauds to some degree, but such detection is possible only after fraudulent impressions and clicks have been created. While this may be feasible for mobile apps, we explore a qualitatively different approach: to detect fraudulent behavior by analyzing the structure of the app, an approach that can detect placement frauds more effectively and before an app is used. Our approach leverages the highly specific, and legally enforceable, terms and conditions that ad networks place on app developers. This project proposes techniques to accurately locate the ranking fraud by mining the active periods, namely leading sessions, of mobile Apps. Proposed approach investigate three types of evidences, i.e., ranking based evidences, rating based evidences and review based evidences, by modeling Apps’ ranking, rating and review behaviors through statistical hypotheses tests. It proposes an optimization based aggregation method. The proposed frame-work is scalable and can be extended with other domain-generated evidences for ranking fraud detection.