A data mining approach in opponent modeling

Authors

Remedios De Dios Bulos, De La Salle University, Manila
Conirose Dulalia, De La Salle University, Manila
Peggy Sharon L. Go, De La Salle University, Manila
Pamela Vianne C. Tan, De La Salle University-Manila
Ma Zaide Ilene O. Uy, De La Salle University, Manila

College

College of Computer Studies

Department/Unit

Software Technology

Document Type

Conference Proceeding

Source Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Volume

3809 LNAI

First Page

993

Last Page

996

Publication Date

1-1-2005

Abstract

In offline opponent modeling, large datasets can be utilized as training data to model the opponent. In the Coach competition of RoboCup Soccer, offline opponent modeling can be adopted to train the coach learn about the opponent's behavior patterns. Data-mining techniques, particularly decision-tree construction can be applied in identifying interesting behavior patterns of the opponent. This research explores the use of the decision-tree algorithm C4.5 to generate classification rules that will embody the offensive and defensive strategies (plans) of the coach against its opponent(s). To achieve this objective, the SimSoccer Coach system is built. © Springer-Verlag Berlin Heidelberg 2005.

html

Digitial Object Identifier (DOI)

10.1007/11589990_127

Recommended Citation

Bulos, R. D., Dulalia, C., Go, P. L., Tan, P. C., & Uy, M. O. (2005). A data mining approach in opponent modeling. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3809 LNAI, 993-996. https://doi.org/10.1007/11589990_127

Disciplines

Computer Sciences

Keywords

Data mining; Soccer—Defense; Soccer—Training

Upload File

wf_yes