Analysis of Content Popularity in Social Bookmarking Systems - Evolving Application Domains of Data Warehousing and Mining - page 257

Yang, Y., & Pedersen, J. O. (1997).A Comparative
Study on Feature Selection in Text Categoriza-
tion. In D. H. Fisher, (Ed.), Proceedings of the
Fourteenth international Conference on Machine
Learning (July 08 - 12, pp. 412-420). San Fran-
cisco: Morgan Kaufmann Publishers Inc.

6

Although the semantics of a digital resource
can be also conveyed by other kinds of
features (e.g. audio, visual), we restrict our
study to the semantics conveyed by textual
features of the content.

7

Whether a story jumps to the 'Popular' sec-
tion or not does not solely depend on the
number of Diggs it receives (although it is
certainly taken into account). The Digg ad-
ministrators make this decision on the basis
of a set of proprietary criteria and heuristics,
which they keep secret since sharing such
knowledge would render the system prone
to malicious attacks (e.g. to artificially boost
the popularity of a story).

Zhang, J.,Ackerman, M. S., &Adamic, L. (2007).
Expertise networks in online communities: struc-
ture and algorithms. In Proceedings of the 16th
international Conference on World Wide Web,
Banff, Alberta, Canada, May 08 - 12, WWW '07,
(pp. 221-230). New York: ACM.

Zhang, Z., & Varadarajan, B. (2006). Utility scor-
ing of product reviews. In Proceedings of the 15th
ACM international Conference on information
and Knowledge Management, Arlington, VA,
November 06 - 11, (CIKM '06, pp. 51-57). New
York: ACM.

8

Only stories with |H r | > 20 were studied to
prevent the 'noisy' time series from distort-
ing the resulting aggregate time series of the
non-popular stories.

9

Users with |R u | < 20 were not considered in
the sample selection in order to prevent users
with sparse (and therefore noisy) activity to
affect the aggregate activity time series.

endnoteS

10

The filtering rules of |R u | > 20 and |H r | >
20 (same as above) were applied here too.

1

An estimate of the allocation of internet users'
time to different online activities is provided
by the Online PublishersAssociation through
the Internet Activity Index (IAI) in: http://
www.online-publishers.org.

11

The post in http://blog.Digg.com/?p=106
provides further evidence in favor of this
speculation.

2

http://www.compete.com

3

http://www.quantcast.com

4

http://www.youtube.com

5

http://ucc.daum.net