Designing a Click Fraud Detection Algorithm - Exposing Suspect Networks

Abstract

As the use of the Web expands and appears almost everywhere in business' practices, new algorithmic problems appear and need to be efficiently handled; one of them that has attracted the attention of both researchers and practitioners is click fraud. Click fraud can be defined as the practice of repetitively clicking on search ads without being actually interested in the content of the related links, with the intention of either increasing the Website's profits or exhausting an advertiser's budget. In this work, we propose an algorithm, which exposes suspect networks instead of single IPs, based on utilizing efficient data structures that have not been employed in previous works.