We examine the challenges behind recommendations in social content sites. We use collaborative
tagging sites (think del.icio.us, YouTube and Yahoo!Travel) as our application and report on our
experiments in harvesting the collective tagging behavior to serve relevant content (think URLs,
videos, travel destinations) to users. We address well-known and lesser-known problems in
recommender systems such as over-specialization and data management for the masses. We
conclude with open questions.
Online reviews are an important asset for users deciding to buy a product, see a movie, or go to a
restaurant, as well as for businesses tracking user feedback. However, most reviews are written in
a free-text format, and are therefore difficult for computer systems to understand, analyze, and
aggregate. One consequence of this lack of structure is that searching text reviews is often frustrating
for users; keyword searches typically do not provide good results as the same keywords routinely
appear in good and in bad reviews. User experience would be greatly improved if the structure and
sentiment information conveyed in the content of the reviews were taken into account. Our work focuses
on identifying this structure and sentiment information from free-text reviews, and using this knowledge
to improve user experience in accessing reviews. Specifically, we focused on improving recommendation
accuracy in a restaurant review scenario.
We report on our classification effort, and on the insight on user-reviewing behavior that we gained
in the process. We propose new ad-hoc and regression-based recommendation measures, that both take into
account the textual component of user reviews. Our results show that using textual information results
in better general or personalized restaurant score predictions than those derived from the numerical star
ratings given by the users.