| Preface | p. ix |
| Introduction to Topic Detection and Tracking | p. 1 |
| Introduction | p. 1 |
| TDT tasks | p. 3 |
| History of TDT | p. 7 |
| TDT 1999 and TDT 2000 | p. 10 |
| The Future of TDT | p. 13 |
| Topic Detection and Tracking Evaluation Overview | p. 17 |
| Introduction | p. 17 |
| TDT Definitions: Stories, Events, and Topics | p. 18 |
| TDT Corpora | p. 19 |
| Evaluation Methodology | p. 20 |
| Task Definitions | p. 25 |
| Summary | p. 30 |
| Corpora for Topic Detection and Tracking | p. 33 |
| Introduction | p. 33 |
| Overview of TDT Corpus Development | p. 35 |
| Collection of Raw Data | p. 36 |
| Transcription | p. 38 |
| Story Segmentation | p. 39 |
| Topic Definition | p. 42 |
| Topic Annotation | p. 45 |
| Corpus Formats | p. 54 |
| Some Properties of the Corpus | p. 61 |
| Conclusion | p. 64 |
| Probabilistic Approaches to Topic Detection and Tracking | p. 67 |
| Introduction | p. 67 |
| Core TDT Technologies | p. 68 |
| Corpus Processing | p. 75 |
| Tracking | p. 75 |
| Detection | p. 77 |
| Crosslingual TDT | p. 80 |
| Conclusions and Future Work | p. 81 |
| Acknowledgments | p. 82 |
| Multi-strategy Learning for TDT | p. 85 |
| Introduction | p. 85 |
| Segmentation | p. 87 |
| Topic and Event Tracking | p. 88 |
| Topic Detection | p. 96 |
| First Story Detection | p. 99 |
| Story Link Detection | p. 101 |
| Multilingual TDT | p. 107 |
| Concluding Remarks | p. 111 |
| Statistical Models of Topical Content | p. 115 |
| Introduction | p. 115 |
| Models of Story Generation | p. 117 |
| Tracking Systems | p. 120 |
| Detection System | p. 128 |
| Summary | p. 132 |
| Segmentation and Detection at IBM | p. 135 |
| Story Segmentation | p. 135 |
| Topic Detection | p. 142 |
| Acknowledgements | p. 147 |
| A Cluster-Based Approach to Broadcast News | p. 149 |
| Introduction | p. 149 |
| Segmentation | p. 152 |
| Detection | p. 154 |
| Tracking | p. 163 |
| Acknowledgements | p. 173 |
| Signal Boosting for Translingual Topic Tracking | p. 175 |
| Introduction | p. 176 |
| The Signal-to-Noise Perspective | p. 177 |
| Topic Tracking System Architecture | p. 178 |
| Contrastive Conditions | p. 184 |
| Conclusions and Future Work | p. 191 |
| Acknowledgments | p. 194 |
| Explorations Within Topic Tracking and Detection | p. 197 |
| Introduction | p. 197 |
| Basic System | p. 198 |
| Tracking | p. 203 |
| Cluster Detection | p. 205 |
| First Story Detection | p. 208 |
| Link Detection | p. 208 |
| Bounds on Effectiveness | p. 216 |
| Automatic Timeline Generation | p. 219 |
| Conclusions | p. 222 |
| Towards a "Universal Dictionary" for Multi-Language IR Applications | p. 225 |
| Introduction | p. 225 |
| Our TDT tracking algorithm | p. 229 |
| The "Universal Dictionary" experiment | p. 236 |
| Conclusions and Directions for Future Work | p. 239 |
| An NLP & IR Approach to Topic Detection | p. 243 |
| Introduction | p. 243 |
| General System Framework | p. 245 |
| Representation of News Stories and Topics | p. 246 |
| Similarity and Interpretation of a Two-threshold Method | p. 248 |
| Multilingual Topic Detection | p. 250 |
| Development Experiments | p. 256 |
| Evaluation | p. 259 |
| Discussion | p. 261 |
| Concluding Remarks and Future Works | p. 262 |
| Index | p. 265 |
| Table of Contents provided by Syndetics. All Rights Reserved. |