Information Technology Reference
In-Depth Information
Chapter 3
User Modeling on Social Multimedia Activity
Abstract
The increasing social multimedia activities conducted on multimedia
sharing web sites reveal user attributes, such as age, gender, and personal interest,
which have been exploited for user modeling, retrieval, and personalization. While
existing user modeling solutions are devoted to inferring user attribute independently,
in this chapter, we investigate the problem of relational user attribute inference. The
task of attribute relation mining and user attribute inference are addressed in a unified
framework.
3.1 Introduction
An intelligent social multimedia service is expected to understand the users' urgent
needs and capture their preferences, so as to push the most interesting multimedia
resources to the most desired users. However, understanding user is not a trivial task.
On one hand, due to the privacy issues, most users are not willing to explicitly provide
their profile and preference information. On the other hand, users' information needs
and preferences are changing over time, making user modeling in a dynamic fashion.
Along with the explosive prevalence of social media networking, more and more
people are engaged in online social media networks. Facebook now reaches 85% of
the world's Internet-using population, i.e., 1.26 billion people in total. Social media
users are creating and sharing large-scale multimedia information, such as textual
posts, photos, and videos. For example, YouTube has reported that in every minute
there are 100h of video uploaded , with more than 6 billion h watch every month.
1
Facebook users have contributed to 250 billion photos totally.
2
Such rich interaction
with multimedia information reveals important clues of user information including
biographic facts (e.g., age, gender, relationship), personal interest (e.g., politics,
technology, entertainment, sports), occupation information (e.g., researcher, student,
software engineer, musicians), and emotional orientation (e.g., optimistic, negative).
