Information Technology Reference
In-Depth Information
5.5 Discussions
The idea of exploiting overlapped users toward cross-network user modeling and
multimedia knowledge association actually opens up possibilities to a very inter-
esting direction. People involve with social multimedia by interacting with het-
erogeneous social multimedia knowledge, e.g., multimedia semantics, geographic
patterns, people consuming patterns, and social interactions. The association among
different social media activities will lead to insightful observations, contribute to col-
lective utilization, and facilitate advanced social media analysis and applications. For
example, the association between user watching activity in
YouTube
and transaction
activity in
Amazon
leads to understanding between user interest and consuming mod-
els, and facilitates cross-network product target advertising. “Multimedia” research
under social media circumstances may understand not only text, image, video, but
the association among heterogeneous social media knowledge.
The user-centric nature of social multimedia inspires us to understand the hetero-
geneous knowledge by “how we experience the world” [
25
]. Instead of conducting
analysis from scratches, the different activities that overlapped users contribute in
different social media networks can be employed as human supervision. This actu-
ally borrows the essence of crowdsourcing where the collective human intelligence
is aggregated.
For cross-network user modeling, we summarize the suggested procedures as:
(1) Determine the user model type (e.g., LBS model and consumer model), identify
the related social media networks, and crawl a dataset of overlapping uses and their
heterogeneous user data. (2) Conduct data analysis on the heterogeneous user data to
find complementary or collaborative characteristics. (3) Develop cross-network per-
sonalized solutions based on the derived data observations. The suggested procedures
for cross-network knowledge association is summarized as: (1) Determine the het-
erogeneous social knowledge involved in different social media networks. (2) Extract
heterogeneous topics on each network and conduct cross-network topic association
based on the observed overlapping users. (3) Design collaborative applications based
on the derived heterogeneous knowledge association.
In the future, in addition to instantiate the suggested procedures with advanced
algorithms and in more practical problems, we will be working toward extending the
overlapped users to connect not only social media networks, but cyber and physical
spaces, e.g., the online and real-world behavior patterns.
References
1. Abel, F., Araújo, S., Gao, Q., Houben. G.-J.: Analyzing cross-system user modeling on the
social web. In: Web Engineering, pp. 28-43. Springer (2011)
2. Althoff, T., Borth, D., Hees, J., Dengel. A.: Analysis and forecasting of trending topics in online
media streams. In: Proceedings of the 21st ACM International Conference on Multimedia,
pp. 907-916 (2013)
