B. Information Processing
A. Information Gathering
C. Incremental Mining
A. Information Gathering
Game Usage Log(s)
(low-level)
Generalized Game
Usage Log(s)
Patterns, Clusters, e.g.
Rules or Statistics
Recommender Service
Massive Multiplayer
Game Service
Figure 3: Game Usage Mining Process
An hierarchical approach of dividing high-
level actions (e.g.
performing a large task)
into several sub-tasks with corresponding
actions have shown useful in the creation of
intelligent Quake monsters [
5
]. This supports
our aggregation approach of actions.
We suggest that the behavior of non-
personal characters (e.g.
monsters) are also
logged and processed in the same way as the
players. This in order to be able to recreate
the sessions and the experience of the player.
These data must be combined with global in-
formation about the game (storyline, major
events, user interface) in the mining process
(part C). Results of the mining process are pat-
terns (e.g. rules or statistics) that can either
be used as input to a recommender service or
as metrics to the game service operator(s).
Acknowledgements
We would like to thank Magnus Lie Hetland for
fruitful discussions about computer games in
general. We would also like to thank Professor
Mihhail Matskin. This work is supported by
the Norwegian Research Council in the frame-
work of the Distributed Information Technol-
ogy Systems (DITS) program and the (
ElCo-
mAg
) project.
6
Conclusion
The contribution of this paper has been
threefold,
first we defined types of data
mining in computer, second we provided a
classification of computer games from a data
mining viewpoint, and third we compared
information gathering in web usage mining
and game usage mining as well as proposing
a common game log format to enable game
usage mining.
Future work include determining the actual
representation of game log files (e.g.
which
attributes and which concrete efficient repre-
sentation), determine what type of usage min-
ing is most useful in massive multiplayer com-
puter games (e.g clustering, classification and
incremental sequence mining [
10
,
7
]), and how
existing web usage mining architectures and
systems can be adapted to a support scalable
game usage mining setting.
References
[1] Ahmed Abdelkhalek, Angelos Bilas, and An-
dreas Moshovos.
Behavior and performance
of interactive multi-player game servers.
In
Proceedings of IEEE International Symposium
on Performance Analysis of Systems and Soft-
ware. IEEE, November 2001.
[2] Robert Cooley,
Bamshad Mobasher,
and
Jaideep Srivastava.
Web mining: Informa-
tion and pattern discovery on the world wide
web. In Proceedings of the 9th IEEE Interna-
tional Conference on Tools with Artificial In-
telligence (ICTAI'97). IEEE, November 1997.
[3] Robert Cooley,
Bamshad Mobasher,
and
Jaideep Srivastava.
Data preparation for
mining world wide web browsing patterns.
Knowledge and Information Systems, 1(1):5
32, February 1999.
[4] Moon Ihlwan. The champs in online games.
Business Week, pages 2728, July 23, 2001.
[5] John Laird. It knows what you're going to do:
Adding anticipation to a quakebot. In Proceed-
ings of the 5th International Conference on
Autonomous Agents. ACM, ACM Press, 2001.
Paper E
105