vassilis papataxiarhis , v.tsetsos, i.karali, p.stamatopoulos, and s.hadjiefthymiades

15
Vassilis Papataxiarhis, V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades [email protected] Department of Informatics and Telecommunications University of Athens – Greece RuleApps-2009, 21 Sep. 2009, Cottbus RuleApps-2009, 21 Sep. 2009, Cottbus i-footman: A Knowledge-Based Framework for Football Managers

Upload: brooke-simon

Post on 04-Jan-2016

45 views

Category:

Documents


0 download

DESCRIPTION

i-footman: A Knowledge-Based Framework for Football Managers. Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades [email protected] Department of Informatics and Telecommunications University of Athens – Greece RuleApps-2009, 21 Sep. 2009, Cottbus. Outline. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Vassilis Papataxiarhis, V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

[email protected]

Department of Informatics and TelecommunicationsUniversity of Athens – Greece

RuleApps-2009, 21 Sep. 2009, CottbusRuleApps-2009, 21 Sep. 2009, Cottbus

i-footman: A Knowledge-Based Framework for Football

Managers

Page 2: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Introduction

Functionality and Provided Services

Application Models and Rules

Implementation

Simulation Results

Conclusions

Outline

Page 3: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

What is i-footman?

◦ A decision support system for football managers

◦ Based on Semantic Web technologies

Main Idea

◦ Provide effective tactical guidelines to face an opponent

Restrictions

◦ Empirical/Subjective Knowledge about football

◦ Lack of statistics and ergometric results

◦ No relevant approach (academic or industrial)

Goals

◦ Model the basic knowledge of the domain

◦ Extensibility (in terms of quality and provided services)

Introduction

Page 4: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Methodology◦ Interview 2 domain experts (i.e. football

managers)

◦ Questionnaires

Knowledge acquisition about:◦ the application domain of football

◦ the desired services

◦ the key features of football players and teams

◦ the tactical guidelines that should be supported by the system

Goal: Incorporate the derived knowledge to the rules and application models

Knowledge Elicitation

Page 5: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

i-footman

Rule Engine

Football Players Ontology

IdentificationFormation

DL-Reasoner

Football TeamsOntology

Rules

Player Selection Tactical Instructions

user

reuses

reusesreuses

i-footman Architecture

Page 6: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Football Players

Ontology

DL-ReasoningRules

Execution

Formation andPlayer Selection

Rules

Identification andTactical

Instructions Rules

Formation

Composition

Strengths/Weaknesses

InstructionsFootball Teams

Ontology

Players Data

Teams Data

Functionality

Page 7: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Ontological Models (1/2)

Expressed in OWL-DL and provide a common vocabulary Football Players Ontology (FPO)

◦ Some metrics: 71 concepts, 43 object prop., 3 datatype prop., each player instance is described by 22 concept inst. and 9 property inst.

◦ It models: Position of players

Technical and physical capabilities

Types of players

E.g., fpo:CreativeMiddlefielder≡ (fpo:hasPassing.GoodAbility ⊔ fpo:hasPassing.VeryGoodAbility) ⊓ fpo:playsInPosition.Middlefielder

Football Teams Ontology (FTO)◦ It models main features and types of teams

Page 8: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Ontological Models (2/2)

Simplified version of FPO Key concepts

◦ Player, Position, PlayerFeature

FTO imports FPO◦ classifies teams according

to the features of its players

◦ models tactical instructions allowing the execution of rules

Page 9: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Expressed in terms of SWRL◦ Motivation: integration of rules and ontologies in the same

logical language Exploit the vocabulary of FPO and FTO Define more complex concepts and relationships Constitute the main part of the knowledge acquired by

interviewing the experts Extensible set of rules Four main categories of rules for the:

◦ identification of team weaknesses/advantages◦ selection of an appropriate tactical formation◦ player selection◦ recommendation of appropriate tactical instructions

Rules

Page 10: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Identification Rule◦ fto:hasStartingPlayer (?t1,?p1) ∧ fto:hasStartingPlayer (?t1,?p2) ∧

fpo:QuickOffensivePlayer (?p1) ∧ fpo:QuickOffensivePlayer (?p2) → fto:dangerousAtCounterAttack (?t1,true).

Formation Rule◦ fto:myTeamPlaysAgainst(?t1,?t2) ∧ fto:TeamWith3CentralDefenders(?

t2) ∧ fto:TeamWith3CentralPlayers(?t2) ∧ fto:TeamWithSideMFs(?t2) ∧ fto:TeamWith2Attackers(?t2) → fto:playsWith3CentralDefenders(?t1, true).

Player Selection Rule◦ fto:myTeamPlaysAgainst(?t1,?t2) ∧ fpo:playsWith1Striker(?t1) ∧

fpo:GoodStriker (?p1) ∧ fpo:isMemberOf(?p1,?t1) → fpo:isSuggestedTo(?p1,?t1).

Tactical Instruction Rule◦ fto:myTeamPlaysAgainst(?t1,?t2) ∧ fto:TeamWithNoBacks(?t2) ∧

fto:TeamWithWingers(?t1) → fto:shouldAttackFromTheWings(?t1, true).

Rules Examples

Page 11: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Web Ontology Language (OWL-DL) Semantic Web Rule Language (SWRL) Pellet Reasoner (v. 1.5.1) Jess Rule Engine Protégé SWRL Jess Tab Protégé OWL API SPARQL Jena2 inference module – Jena API Apache Tomcat

Implementation Details

Page 12: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Simulation of football matches in 2 platforms with and without the intervention of i-footman

2 Scenarios◦ Teams with similar ratings

◦ i-footman controls a weaker team

40 games in each platform (80 games in total) Scenario 1

Evaluation (1/2)

Barcelona FC vs. Real Madrid FC (match results)

14

10

1615

17

8

02

46

810

1214

1618

Wins Draws Losses

CPU

i-footman

Barcelona FC vs. Real Madrid FC (goals)

45

51

61

36

0

10

20

30

40

50

60

70

Goals + Goals -

CPU

i-footman

Page 13: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Scenario 2

No significant improvement when controlling a better team

Performance Evaluation

Evaluation (2/2)

Average Response Time = 7740ms

Olympiacos SFP vs. Real Madrid FC (match results)

3

11

26

5

1817

0

5

10

15

20

25

30

Wins Draws Losses

CPU

i-footman

Olympiacos SFP vs. Real Madrid FC (goals)

23

31

87

62

0

10

20

30

40

50

60

70

80

90

100

Goals + Goals -

CPU

i-footman

Page 14: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Contributions◦ A knowledge-based system based on SW technologies◦ An extensible framework for football managers◦ FPO, FTO ontologies

Open Issues◦ Integrated reasoning module for

handling rules and ontologiesseamlessly

◦ Real data are not available Future Work

◦ Automated ontology creation by statistics and ergometric data◦ Learning rules by historical data stemmed from simulations

without the intervention of i-footman◦ Adoption of fuzzy approaches to deal with uncertainty

Conclusion

OntologicalReasoning

InferredKnowledge

RulesExecution

InferredKnowledge

Page 15: Vassilis Papataxiarhis , V.Tsetsos, I.Karali, P.Stamatopoulos, and S.Hadjiefthymiades

Thank you!Thank you!

http://www.di.uoa.gr/~vpap/i-footman