Macintosh / Ellis / Allen Applications and Innovations in Intelligent Systems XII

Proceedings of AI-2004, the Twenty-fourth SGAI International Conference on Innhovative Techniques and Applications of Artificial Intelligence
1. Auflage 2007
ISBN: 978-1-84628-103-7
Verlag: Springer-Verlag
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)

E-Book, Englisch, 279 Seiten

A. L. Macintosh, Napier University, UK The papers in this volume are the refereed application papers presented at ES2004, the Twenty-fourth SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence, held in Cambridge in December 2004. The conference was organised by SGAI, the British Computer Society Specialist Group on Artificial Intelligence. This volume contains twenty refereed papers which present the innovative application of a range of AI techniques in a number of subject domains. This year, the papers are divided into sections on Synthesis and Prediction, Scheduling and Search, Diagnosis and Monitoring, Classification and Design, and Analysis and Evaluation This year's prize for the best refereed application paper, which is being sponsored by the Department of Trade and Industry, was won by a paper entitled 'A Case-Based Technique for Tracking Concept Drift in Spam Filtering'. The authors are Sarah Jane Delany, from the Dublin Institute of Technology, Ireland, and Padraig Cunningham, Alexey Tsymbal, and Lorcan Coyle from Trinity College Dublin, Ireland. This is the twelfth volume in the Applications and Innovations series. The Technical Stream papers are published as a companion volume under the title Research and Development in Intelligent Systems XXI. On behalf of the conference organising committee I should like to thank all those who contributed to the organisation of this year's application programme, in particular the programme committee members, the executive programme committee and our administrators Linsay Turbert and Collette Jackson.

Autoren/Hrsg.

Macintosh, Ann

Ellis, Richard

Allen, Tony

Inhaltsverzeichnis

1;APPLICATION PROGRAMME CHAIR'S INTRODUCTION;5
2;ACKNOWLEDGEMENTS;6
3;APPLICATIONS PROGRAMME COMMITTEE;7
4;CONTENTS;8
5;BEST APPLICATION PAPER;11
5.1;A Case-Based Technique for Tracking Concept Drift in Spam Filtering;12
5.1.1;1 Introduction;12
5.1.2;2 Spam Filtering and Machine Learning;13
5.1.3;3 The Problem of Concept Drift;14
5.1.3.1;3.1 Definitions and Types of Concept Drift;14
5.1.3.2;3.2 Approaches to Handling Concept Drift;14
5.1.4;4 A Case-Based Approach to Concept Drift;15
5.1.4.1;4.1 Feature Selection;16
5.1.4.2;4.2 Case Retrieval;16
5.1.4.3;4.3 Case-base Management;17
5.1.5;5 Evaluation;18
5.1.5.1;5.1 Experimental Setup;18
5.1.5.2;5.2 Evaluation Metrics;18
5.1.5.3;5.3 CBR vs. Naive Bayes;19
5.1.5.4;5.4 Level 1 Learning - Continuous Updating with new Instances;20
5.1.5.5;5.5 Level 2 Learning - Model Rebuild with Feature Reselection;21
5.1.6;6 Conclusions;23
5.1.7;References;23
6;SESSION 1 : SYNTHESIS AND PREDICTION;26
6.1;Matching and Predicting Crimes;27
6.1.1;1. Introduction;27
6.1.2;2. Data and Initial Analysis;28
6.1.3;3. Matching Crimes;30
6.1.3.1;3.1 Single Crimes with Similarity-Based Retrieval;30
6.1.3.2;3.2 Naive Bayes Classifier and Spatio-Temporal Features;31
6.1.3.3;3.3 Matching Series of Crimes with Logic;33
6.1.4;4. Predicting Crimes;35
6.1.4.1;4.1 Survival/Failure Time Analysis;35
6.1.4.2;4.2 Bayesian Network;37
6.1.5;5. Conclusions and Future Directions;38
6.1.6;Acknowledgements;39
6.1.7;References;39
6.2;Story Plot Generation based on CBR;41
6.2.1;1 Introduction;41
6.2.2;2 Related Work;42
6.2.2.1;2.1 Computational Plot Generation;42
6.2.2.2;2.2 Natural Language Generation;42
6.2.3;3 A CBR System for Composing Plot Plans;43
6.2.3.1;3.1 The System Knowledge;43
6.2.3.2;3.2 The CBR Module;46
6.2.4;4 The Natural Language Generation Module;48
6.2.4.1;4.1 Input Data for the NLG Module;48
6.2.4.2;4.2 Stages of the NLG Module;50
6.2.5;5 Conclusions;52
6.2.6;Acknowledgements;53
6.2.7;References;53
6.3;Studying Continuous Improvement from a Knowledge Perspective;55
6.3.1;1. Introduction;55
6.3.1.1;1.1 Bacl^round for the Study;56
6.3.1.2;1.2 Why Apply Knowledge Structure Mapping to Continuous Improvement?;56
6.3.2;2. The Knowledge Structure Mapping;57
6.3.2.1;2.1 Outputs from a KSM Investigation;57
6.3.2.2;2.2 The Value of a Knowledge Centred Approach;59
6.3.3;3* The CI Investigation;60
6.3.3.1;3.1 Defining the Study Focus;60
6.3.3.2;3.2 Forming the Team;61
6.3.3.3;3.3 The Study Approach;62
6.3.3.4;3.4 Analysing and Evaluating Results;63
6.3.3.5;3.5 Report Delivery;63
6.3.4;4. Project Results;64
6.3.4.1;4.1 The Results;64
6.3.4.2;4.2 Analysis of the Results;65
6.3.4.3;4.3 Observations and Recommendations;67
6.3.5;5. Initial Business Reaction;69
6.3.6;6. Conclusion;70
6.3.7;References;71
6.4;ReTAX+: A Cooperative Taxonomy Revision Tool;72
6.4.1;1. Introduction;72
6.4.2;2. Related Work;73
6.4.3;3. System Architecture;74
6.4.4;4. Definitions for Ontological Analysis;75
6.4.4.1;4.1 Taxonomy;75
6.4.4.2;4.2 Class & Attributes;75
6.4.5;5. Operations on Taxonomy;76
6.4.5.1;5.1 Maintenance of x and dri;76
6.4.5.2;5.2 Manipulation of Attributes;77
6.4.5.3;5.3 Manipulation of Classes;78
6.4.5.4;5.4 Refinement Strategies;81
6.4.6;6. Implementation;83
6.4.6.1;6.1 Planned Enhancements;83
6.4.7;7. Conclusion;84
7;SESSION 2: SCHEDULING AND SEARCH;86
7.1;A Non-Binary Constraint Ordering Approach to Scheduling Problems;87
7.1.1;1 Introduction;87
7.1.2;2 Definition and Algorithms;89
7.1.2.1;2.1 Variable Ordering;89
7.1.2.2;2.2 Value Ordering;90
7.1.3;3 Example of Non-binary Scheduling Problem;90
7.1.4;4 Constraint Ordering Heuristic (COH);92
7.1.4.1;4.1 Specification of COH;93
7.1.5;5 Analysis of COH;95
7.1.6;6 Evaluation of COH;95
7.1.7;7 Conclusion and future work;99
7.1.8;References;99
7.2;A Heuristic Based System for Generation of Shifts with Breaks;101
7.2.1;1 INTRODUCTION;101
7.2.2;2 PROBLEM DESCRIPTION;103
7.2.3;3 APPLICATION DESCRIPTION;104
7.2.3.1;3.1 Generation of breaks;104
7.2.4;4 APPLICATION USE;106
7.2.4.1;4.1 Definition of temporal requirements;107
7.2.4.2;4.2 Constraints regarding shift types;107
7.2.4.3;4.3 Weights of criteria;109
7.2.4.4;4.4 Definition of break types;109
7.2.4.5;4.5 Generation of shifts;110
7.2.5;5 CONCLUSION;110
7.2.6;Acknowledgment;112
7.2.7;References;112
7.3;A Topological Model Based on Railway Capacity to Manage Periodic Train Scheduling;113
7.3.1;1 Introduction;113
7.3.2;2 Problem Topology;115
7.3.2.1;2.1 Railway Traffic Rules, topological and requirement constraints;116
7.3.2.2;2.2 General System Architecture;117
7.3.3;3 Railway Capacity;118
7.3.4;4 Topological Constraint Optimization Technique;120
7.3.4.1;4.1 Topological Technique;121
7.3.5;5 Evaluation;123
7.3.6;6 Conclusions;125
7.3.7;References;126
7.4;Collaborative Search; Deployment Experiences;127
7.4.1;1 Introduction;127
7.4.2;2 Repetition &: Regularity in Web Search;128
7.4.3;3 A Review of Collaborative Search;129
7.4.4;4 I-SPY;130
7.4.4.1;4.1 Creating a Community;131
7.4.4.2;4.2 Example Session;132
7.4.5;5 Live-User Evaluation;133
7.4.5.1;5.1 Methodology;134
7.4.5.2;5.2 Test Scores;134
7.4.5.3;5.3 Group Differences;136
7.4.5.4;5.4 Result Selections;137
7.4.5.5;5.5 Summary;138
7.4.6;6 Lessons Learned;138
7.4.6.1;6.1 Exploiting Query Similarity;138
7.4.6.2;6.2 Query Recommendation;139
7.4.6.3;6.3 Result Diversity;139
7.4.7;7 Conclusions;139
7.4.8;References;140
8;SESSION 3: DIAGNOSIS AND MONITORING;141
8.1;A Model-Based Approach to Robot Fault Diagnosis;142
8.1.1;1 Introduction;142
8.1.2;2 Problem Formulation and Solution;143
8.1.2.1;2.1 Robotic Fault Description;143
8.1.2.2;2.2 Robot Diagnosis System;144
8.1.3;3 Robotic Fault Reasoning;144
8.1.3.1;3.1 Robotic Bounds Generation;145
8.1.3.2;3.2 Interval Filter;146
8.1.3.3;3.3 Component-Based Robotic Reasoning;147
8.1.4;4 Case study;150
8.1.4.1;4.1 Simulation results;150
8.1.4.2;4.2 Fault report;152
8.1.5;5 Conclusion;152
8.1.6;Acknowledgements;153
8.1.7;Notation;153
8.1.8;References;154
8.2;Automating the Analysis and Management of Power System Data using Multi-agent Systems Technology;156
8.2.1;1. Introduction;156
8.2.2;2. Post-fault Analysis and Management of Power System Data related to Protection Operation;157
8.2.2.1;2.1 Monitoring Systems and Data sources;158
8.2.2.2;2.2 Post-fault Analysis using Existing Software Tools;161
8.2.3;3. Protection Engineering Diagnostic Agents;163
8.2.3.1;3.1 Agents and their Interactions;163
8.2.4;4. Industrial Implementation of PEDA;165
8.2.4.1;4.1 Issues arising during industrial implementation;165
8.2.4.2;4.2 Resolving issues related to industrial implementation;166
8.2.4.3;4.3 Agents and their interactions;167
8.2.5;5. Concluding Remarks;169
8.2.6;References;169
8.3;The Industrialisation of a Multi-Agent System for Power Transformer Condition Monitoring;170
8.3.1;1 Introduction;170
8.3.2;2 Transformer Condition Monitoring;171
8.3.3;3 COMMAS;172
8.3.3.1;3.1 Functional Design;172
8.3.3.2;3.2 Detailed Design;174
8.3.4;4 Aspects of Extensible Design;176
8.3.4.1;4.1 The Java Agent DEvelopment Framework (JADE);177
8.3.4.2;4.2 Ontology and Content Language;177
8.3.4.3;4.3 Raw Data Formats;179
8.3.5;5 Running COMMAS;181
8.3.6;6 Conclusion;181
8.3.7;References;182
9;SESSION 4: CLASSIFICATION AND DESIGN;184
9.1;An Improved Genetic Programming Technique for the Classification of Raman Spectra;185
9.1.1;1 Introduction;185
9.1.2;2 Description of Task;186
9.1.3;3 Analysis Techniques;187
9.1.4;4 Experimental Results;191
9.1.5;5 Conclusions & Future Work;195
9.2;PROPOSE - Intelligent Advisory System for supporting Redesign;197
9.2.1;1. Introduction;197
9.2.2;2. Analysis-Based Optimisation in Design Process;198
9.2.3;3. Development of the Intelligent Redesign System;200
9.2.4;4. The Knowledge Base;201
9.2.5;5. The Shell of the System;203
9.2.6;6. Evaluation of the System;206
9.2.7;Conclusions;210
9.2.8;Acknowledgements;210
9.2.9;References;211
9.3;The Designers' Workbench: Using Ontologies and Constraints for Configuration;213
9.3.1;1. Backround;213
9.3.2;2. Aims of Designers' Workbench;214
9.3.3;3. Related work;215
9.3.3.1;3.1 Constraint-based approaches;215
9.3.3.2;3.2 Ontology-based approaches;216
9.3.3.3;3.3 Combining constraints and ontologies;216
9.3.4;4. An illustrative example;217
9.3.5;5. Functionality;219
9.3.6;6. Additional features of the Designers* Workbench;220
9.3.6.1;6.1 Graph-based display of configuration;221
9.3.6.2;6.2 Checking incomplete configurations;221
9.3.6.3;6.3 Constraint rationales;221
9.3.7;7. Preliminary evaluation;222
9.3.8;8. Future work;222
9.3.8.1;8.1 Propagation of values;222
9.3.8.2;8.2 Functional descriptions;223
9.3.8.3;8.3 Design rationale storage (argumentation);223
9.3.8.4;8.4 Constraint input;223
9.3.8.5;8.5 Ontology change;223
9.3.9;9. Summary;224
9.3.10;10. Acknowledgements;224
9.4;A Visualisation Tool to Explain Case-Base Reasoning Solutions for Tablet Formulation;226
9.4.1;1 Introduction;226
9.4.2;2 Solutions Require Explanations;228
9.4.3;3 What Needs to be Explained;229
9.4.3.1;3.1 Knowledge Containers;229
9.4.3.2;3.2 The Proposed Solution;230
9.4.4;4 Textual FORMUCASE;230
9.4.5;5 FORMUCASEVIZ;231
9.4.5.1;5.1 Ordering the Attributes;232
9.4.6;6 User Evaluation of FormuCaseViz;234
9.4.7;7 Related Work;236
9.4.8;8 Conclusions and Future Work;237
9.4.9;Acknowledgments;237
9.4.10;References;237
10;SESSION 5: ANALYSIS AND EVALUATION;239
10.1;Formal Analysis of Empirical Traces in Incident Management;240
10.1.1;1. Introduction;240
10.1.2;2. The Hercules Disaster;241
10.1.3;3. The Dakota Disaster;243
10.1.4;4. Categorisation of Error Types;244
10.1.5;5. Modelling Approach;246
10.1.6;6. Formalisation of an Empirical Trace;247
10.1.7;7. Validation of a Trace;249
10.1.8;8. Discussion;252
10.1.9;References;252
10.2;Modelling Expertise for Structure Elucidation in Organic Chemistry Using Bayesian Networks;254
10.2.1;1 Introduction;254
10.2.2;2 Related Work;255
10.2.3;3 Bayesian Networks;256
10.2.4;4 Modelling;258
10.2.4.1;4.1 Spectral and Structural Subunits;258
10.2.4.2;4.2 Aromatic Carbons;260
10.2.4.3;4.3 Laws of the Domain;261
10.2.4.4;4.4 Assignment of Probability Values;263
10.2.5;5 Results and Outlook;264
10.2.6;References;266
10.3;Evaluation of a Mixed-Initiative Dialogue Multimodal Interface;268
10.3.1;1. Introduction;268
10.3.2;2. System Architecture;269
10.3.3;3. Operation Overview;270
10.3.4;4. Comparison between the Mixed-initiative and Directed-Dialogue Interfaces;272
10.3.5;5. Accuracy and Usability Test;273
10.3.5.1;5.1 Participants Selection;273
10.3.5.2;5.2 Experimental Setup;274
10.3.5.3;5 3 Statistics Methods for Test Results;274
10.3.5.4;5.4 Speech Recognition Accuracy Result;275
10.3.5.5;5.5 Usability Attributes Result;276
10.3.5.6;5.6 Results according to user age;277
10.3.5.7;5.7 Error Recovery Test Result;278
10.3.6;6. Conclusion and Discussion;279
10.3.7;Acknowledgements;280
10.3.8;References;281
11;AUTHOR INDEX;282