Date of Award:

5-2016

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Computer Science

Committee Chair(s)

Kyumin Lee

Committee

Kyumin Lee

Committee

Curtis Dyreson

Committee

Tung Nguyen

Abstract

As billions of users have used online social networks such as Facebook, Twitter, and Instagram, many online services have begun adapting signals of the popularity. For example, search engines rank popular users posts in the top results. A post of a popular user is automatically delivered to the followers or likers. Popular companies in online social networks also get good reputation from people. The popularity on Facebook is measured based on the number of likes in a page. Unfortunately, some crowdsourcing websites (e.g.Microworkers.com) or supply-driven marketplaces (e.g.Fiverr.com) began offering fake liking services in which these crowd workers sent fake likes to Facebook page owners and earned money from the page owners. This manipulation has been degrading information trust and threatening trustworthiness of the online social network. To stop these fake liking activities, in this thesis, we (i) collected profiles of fake likers and legitimate users from Facebook, (ii) analyzed characteristics of fake likers and legitimate users, and (iii) proposed and developed fake liker detection approaches. Our experimental results show that our approaches effectively identified fake likers, achieving 87.1% accuracy and 0.1 false positive rate and 0.14 false negative rate.

Checksum

e30f90af4a0c38709fbdbed493e78438

Share

COinS