Font Size: a A A

Online fraud economy: Characterization and defens

Posted on:2018-10-27Degree:Ph.DType:Dissertation
University:Indiana UniversityCandidate:Bulakh, VladFull Text:PDF
GTID:1449390005958268Subject:Computer Science
Abstract/Summary:
More than three billion people are now connected to the Internet around the world. They manage their finances using online banking, purchase goods using e-commerce websites, entertain themselves by watching movies, news, and video clips online, and stay connected with their friends using Online Social Networks (OSNs). Dangers await them as miscreants are exploiting the heavy use of the Internet the best they can. This dissertation characterizes a few of these dangers and proposes defenses. As people shop online and in brick-and-mortar stores, their credit card information can be stolen and sold on underground markets. Our first contribution in this dissertation is a measurement study of three popular marketplaces that sell stolen credit/debit card information online. We find that these marketplaces are thriving, with revenues ranging between $181K-277K per month in 2015. Carding shops are only one of the vectors of a flourishing underground fraud economy. To gain a better perspective on this economy, for our second contribution, we investigate how cracking tools used to test stolen user credentials for various websites operate and develop classifiers capable of detecting packets generated by them. Further, as Internet users check their email, use online banking, and engage in discussions on OSNs, they can fall prey to phishing. As our next contribution, we study phishing websites to understand characteristics that the phished brands can use to proactively defend themselves. Unlike existing approaches, which find phishing websites without regard for who is being phished, we leverage unique advantage that brands have and discover heuristics that would allow brands to detect as much as 86% of phishing attacks against them. For our final contribution, we engage in the fraudulent video economy for a popular video sharing website. We characterize misbehaving videos and user profiles and develop machine learning algorithms that can detect them.
Keywords/Search Tags:Online, Economy
Related items