Fraud Detection for Online Advertising

With the increasing popularity of Internet, online advertising becomes a new marketing opportunity by instant globally advertisements (ads). At the basis, the process of online advertising can be considered as a buyer/seller relationship, where the two of the key participants are publishers (i.e. seller) and advertisers (i.e. buyer). Publishers make money through hosting websites with advertisements, while advertisers pay for having their ads displayed on publishers' websites. However, since “how much the buyer is willing to pay” depends on the number of clicks to the ads, some dishonest publishers attempt to generate invalid traffic to make more money. Publisher fraud is a severer problem for advertisers. Fraud detection introduces methods and algorithms as countermeasures to different attacking mechanisms. This project aims to apply theoretical analysis and state of the art fraud detection mechanisms to increase the efficiency of the detection and significantly reduce the cost of the partner organization.

Faculty Supervisor:

Dr. Wenying Feng

Student:

Qinglei Zhang

Partner:

EQ Advertising Group Ltd.

Discipline:

Computer science

Sector:

Media and communications

University:

Trent University