[PAST EVENT] Haitao Xu, Computer Science - Ph.D. Dissertation Defense

December 1, 2015
12:30pm - 2:30pm
Blow Memorial Hall, Room 201
262 Richmond Rd
Williamsburg, VA 23185Map this location

The continuous expansion of the Internet in the past 20 years has greatly facilitated the booming development of Internet business. Unfortunately, some unscrupulous participants in Internet business conduct fraudulent activities for their own profits at the expense of other parties. In this dissertation, we present our study on the fraudulent activities in two kinds of major Internet businesses - online advertising and E-commerce.

Online advertising is leveraged by online advertisers to deliver marketing messages to potential customers. It serves as a significant source of revenue for web-based businesses and is crucial to a thriving Internet ecosystem. However, click fraud is posing a serious threat to online advertising systems. As the direct victims, advertisers still lack effective defense against click fraud. In this dissertation, we present a novel approach for advertisers to detect click fraud without the helps from ad networks or publishers. Our proposed defense is effective in identifying both clickbots and human clickers, while incurring negligible overhead at both the server and client sides.

In an E-commerce market, a store's reputation is closely tied to its profitability. Sellers' desire to quickly achieve high reputation has fueled a profitable underground business, termed by us as a seller-reputation-escalation (SRE) market. An SRE market operates as a specialized crowdsourcing marketplace and facilitates online sellers to harness human laborers to conduct fake transactions for improving their stores' reputations. In this dissertation, we characterize the SRE markets in terms of its prevalence, business model, market size, and the sellers and laborers involved. We also evaluate the effectiveness of the SRE services on reputation escalation.

Web traffic generated by web bots is contributing to a large proportion of all web traffic on the Internet. E-commerce sites suffer from web bot traffic greatly. In this dissertation, we aim to identify the behavioral patterns of web bots and infer their intents. To this end, we dissected one-month web traffic generated by 99,089 web bots to a large e-commerce site. We found that web bots present unique and different behavioral patterns and preferences than normal logged-on users in terms of active time, search queries, clicks, visited items and stores.

Haitao Xu is a Ph.D. candidate at William & Mary and supervised by Dr. Haining Wang. His dissertation work focuses on investigating fraudulent activities in online business. His research interests lie in web security and online fraud detection. Haitao Xu received his B.S. degree in Computer Science from Zhengzhou University, Zhengzhou, China in 2007, and M.S. in Computer Science from University of Chinese Academy of Sciences, Beijing, China in 2010.

[[vlthompsondopp, Vicki Thompson Dopp]]