Evidence fusion for real time click fraud detection and prevention

Chamila Walgampaya, University of Louisville
Mehmed Kantardzic
Roman Yampolskiy, University of Louisville


From the viewpoint of Dempster-Shafer evidence theory, information obtained from different sources can be considered as pieces of evidence, and as such, multi-sensor based CCFDP (Collaborative Click Fraud Detection and Prevention) system can be viewed as a problem of evidence fusion. In this paper we detail the multi level data fusion mechanism used in CCFDP for real time click fraud detection and prevention. Prevention mechanisms are based on blocking suspicious traffic by IP, referrer, city, country, ISP, etc. Our system maintains an online database of these suspicious parameters. We have tested the system with real world data from an actual ad campaign where the results show that use of multi-level data fusion improves the quality of click fraud analysis. © 2011 Springer Science+Business Media, LLC.