EXPLOITING USER COMMENTS FOR WEB APPLICATIONS

Abstract

The hallmark of the Web 2.0 era is the incorporation of the user in playing a central role, generating content such as comments, tags, votes and other social actions. As a result, the Web has experienced a renaissance in the form of user-generated content. This characteristic of the new Web has presented new challenges in retrieving, managing and utilizing the great volume of online resources.

In this thesis, we leverage online user comments (or interchangeably ?reviews?) ? with the goal of exploiting the rich signals, such as the presence of temporal dynamics, social influence and user preference ? for improving relevant Web applications. We investigate how applications can benefit from leveraging comments, and propose specific techniques for three target applications, namely 1) popularity prediction, 2) item clustering and 3) personalized recommendation. Our investigations into these three applications form the organization of this thesis.