Click Fraud Detection on the Advertiser Side

摘要: Click fraud--malicious clicks at the expense of pay-per-click advertisers--is posing a serious threat to Internet economy. Although click fraud has attracted much attention from security community, as direct victims fraud, advertisers still lack effective defense detect independently. In this paper, we propose novel approach for frauds and evaluate return on investment ROI their ad campaigns without helps networks or publishers. Our key idea is proactively test if visiting clients are full-fledged modern browsers passively scrutinize user engagement. particular, introduce new functionality develop an extensive characterization detection can significantly raise bar committing transparent users. Moreover, our requires little effort be deployed advertiser side. To validate effectiveness approach, implement prototype deploy it large production website; then run 10-day website major network. The experimental results show that proposed in identifying both clickbots human clickers, while incurring negligible overhead server client sides.