Anomalous Behavior Analysis in Social Networks and Consumer Review Websites

Web and social media have been influencing every aspect of today's world, rendering a tremendous amount of data that requires new insights to know about the current society. The usage and dependence on these media has led to their active use by a small yet powerful group of users to sway the sentiment of people for selfish gains. To check the infiltration of these anomalous users, we face two challenges: (1) studying opinions to learn their behaviors, and (2) detecting opinion spam to reduce their effects. We study the behaviors of anomalous users in reviews collected on Yelp, Amazon and social data from Twitter. Using Yelp reviews we explore the temporal behaviors of spammers. Social spammers easily penetrate and are difficult to filter as they adapt to changing filtering algorithms. Using a Twitter dataset, we study their behaviors of success rate, fraudulence, and content posting activities. We uncover that successful spammers have a stronger friendship base and post an amalgam of spam and non-spam contents. We exploit the behaviors learned from Yelp and Twitter to generate spam detection algorithms. Our novel temporal features are instrumental in spam detection in consumer reivews, performing better than existing state-of-the-art approaches. We combine the content-based features and graph based approach embodying social relationships for spam detection in Twitter. Biased random walks and language models significantly improve the classification. We further characterize the review system in Amazon, a leading online marketplace. We use verified purchases as a popularity index to evaluate models of popularity prediction. We find that it is indeed possible to analyze behaviors and develop methods that perform well for anomaly detection in the web, even in these challenging situations.

Name thesis:degree_name

Doctor of Philosophy

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Houston

Year dc:date.issued

2018

Author dc:creator

• K. C, Santosh 1991-

Advisor dc:contributor.advisor

• Mukherjee, Arjun

Committee members dc:contributor.committeemember

• Verma, Rakesh M.
• Gnawali, Omprakash
• Bronk, Chris

• Spam Detection
• Anomaly Detection
• Machine learning
• Deep learning

Statement dc:rights

• The author of this work is the copyright owner. UH Libraries and the Texas Digital Library have their permission to store and provide access to this work. UH Libraries has secured permission to reproduce any and all previously published materials contained in the work. Further transmission, reproduction, or presentation of this work is prohibited except with permission of the author(s).

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/10657/5703

OAI identifier oai:identifier

oai:uh-ir.tdl.org:10657/5703