Using online learning to analyze the opponent's behavior

Ubbo E Visser, Hans Georg Weland

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

Analyzing opponent teams has been established within the simulation league for a number of years. However, most of the analyzing methods are only available off-line. Last year we introduced a new idea which uses a time series-based decision tree induction to generate rules on-line. This paper follows that idea and introduces the approach in detail. We implemented this approach as a library function and are therefore able to use on-line coaches of various teams in order to test the method. The tests are based on two 'models': (a) the behavior of a goal-keeper, and (b) the pass behavior of the opponent players. The approach generates propositional rules (first rules after 1000 cycles) which have to be pruned and interpreted in order to use this new knowledge for one's own team. We discuss the outcome of the tests in detail and conclude that on-line learning despite of the lack of time is not only possible but can become an effective method for one's own team.

Original languageEnglish (US)
Title of host publicationLecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)
EditorsG.A. Kaminka, P.U. Lima, R. Rojas
Pages78-93
Number of pages16
Volume2752
StatePublished - 2003
Externally publishedYes
Event6th Robot World Cup Soccer and Rescue Competitions and Conference - RoboCup 2002 - Fukuoka, Japan
Duration: Jun 19 2002Jun 25 2002

Other

Other6th Robot World Cup Soccer and Rescue Competitions and Conference - RoboCup 2002
CountryJapan
CityFukuoka
Period6/19/026/25/02

Fingerprint

Online Learning
Decision trees
Time series
Decision tree
Proof by induction
Cycle
Line
Simulation
Model

ASJC Scopus subject areas

  • Hardware and Architecture
  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Visser, U. E., & Weland, H. G. (2003). Using online learning to analyze the opponent's behavior. In G. A. Kaminka, P. U. Lima, & R. Rojas (Eds.), Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science) (Vol. 2752, pp. 78-93)

Using online learning to analyze the opponent's behavior. / Visser, Ubbo E; Weland, Hans Georg.

Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). ed. / G.A. Kaminka; P.U. Lima; R. Rojas. Vol. 2752 2003. p. 78-93.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Visser, UE & Weland, HG 2003, Using online learning to analyze the opponent's behavior. in GA Kaminka, PU Lima & R Rojas (eds), Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). vol. 2752, pp. 78-93, 6th Robot World Cup Soccer and Rescue Competitions and Conference - RoboCup 2002, Fukuoka, Japan, 6/19/02.
Visser UE, Weland HG. Using online learning to analyze the opponent's behavior. In Kaminka GA, Lima PU, Rojas R, editors, Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). Vol. 2752. 2003. p. 78-93
Visser, Ubbo E ; Weland, Hans Georg. / Using online learning to analyze the opponent's behavior. Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science). editor / G.A. Kaminka ; P.U. Lima ; R. Rojas. Vol. 2752 2003. pp. 78-93
@inproceedings{29851a5d10854707aa9ce828b3c8f3df,
title = "Using online learning to analyze the opponent's behavior",
abstract = "Analyzing opponent teams has been established within the simulation league for a number of years. However, most of the analyzing methods are only available off-line. Last year we introduced a new idea which uses a time series-based decision tree induction to generate rules on-line. This paper follows that idea and introduces the approach in detail. We implemented this approach as a library function and are therefore able to use on-line coaches of various teams in order to test the method. The tests are based on two 'models': (a) the behavior of a goal-keeper, and (b) the pass behavior of the opponent players. The approach generates propositional rules (first rules after 1000 cycles) which have to be pruned and interpreted in order to use this new knowledge for one's own team. We discuss the outcome of the tests in detail and conclude that on-line learning despite of the lack of time is not only possible but can become an effective method for one's own team.",
author = "Visser, {Ubbo E} and Weland, {Hans Georg}",
year = "2003",
language = "English (US)",
volume = "2752",
pages = "78--93",
editor = "G.A. Kaminka and P.U. Lima and R. Rojas",
booktitle = "Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)",

}

TY - GEN

T1 - Using online learning to analyze the opponent's behavior

AU - Visser, Ubbo E

AU - Weland, Hans Georg

PY - 2003

Y1 - 2003

N2 - Analyzing opponent teams has been established within the simulation league for a number of years. However, most of the analyzing methods are only available off-line. Last year we introduced a new idea which uses a time series-based decision tree induction to generate rules on-line. This paper follows that idea and introduces the approach in detail. We implemented this approach as a library function and are therefore able to use on-line coaches of various teams in order to test the method. The tests are based on two 'models': (a) the behavior of a goal-keeper, and (b) the pass behavior of the opponent players. The approach generates propositional rules (first rules after 1000 cycles) which have to be pruned and interpreted in order to use this new knowledge for one's own team. We discuss the outcome of the tests in detail and conclude that on-line learning despite of the lack of time is not only possible but can become an effective method for one's own team.

AB - Analyzing opponent teams has been established within the simulation league for a number of years. However, most of the analyzing methods are only available off-line. Last year we introduced a new idea which uses a time series-based decision tree induction to generate rules on-line. This paper follows that idea and introduces the approach in detail. We implemented this approach as a library function and are therefore able to use on-line coaches of various teams in order to test the method. The tests are based on two 'models': (a) the behavior of a goal-keeper, and (b) the pass behavior of the opponent players. The approach generates propositional rules (first rules after 1000 cycles) which have to be pruned and interpreted in order to use this new knowledge for one's own team. We discuss the outcome of the tests in detail and conclude that on-line learning despite of the lack of time is not only possible but can become an effective method for one's own team.

UR - http://www.scopus.com/inward/record.url?scp=23144438478&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=23144438478&partnerID=8YFLogxK

M3 - Conference contribution

VL - 2752

SP - 78

EP - 93

BT - Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)

A2 - Kaminka, G.A.

A2 - Lima, P.U.

A2 - Rojas, R.

ER -