Detection of Coalition Attacks for Online Advertising

As the Internet has become part of people’s daily life, online advertising has also been more and more popular for its instant global communications. Of a number of factors on the cost of the advertising, a major consideration is the number of views of the ads, clicks to the ads, and conversion generated by the ads. Publisher fraud is the action that generates invalid traffic to the ads by inflating the number of views, clicks and conversions. With the increasing popularity of online advertising, publisher frauds become more and more sophisticated nowadays, and more advanced techniques are required to detect them. Coalition attacks is a type of recently prevalent publisher frauds that classical detection techniques are insufficient to identify them. The objective of the project is to study and develop new algorithms to automatically detect the coalition fraud actions of the publishers and therefore reduce unnecessary cost of the advertisers. The results will directly benefit the partner organization by increasing the efficiency of their platform to advertisers.

Faculty Supervisor:

Wenying Feng

Student:

Partner:

EQ Advertising Group Ltd

Discipline:

Computer science

Sector:

Information and cultural industries

University:

Trent University

Program: