Mining Text Data

Contains topics across social networks and data mining. This title focuses on Text Embedded with Heterogeneous and Multimedia Data. Text mining applications have experienced tremendous advances because of web 2.0 and social networking applications. Recent advances in hardware and software technology have lead to a number of unique scenarios where text mining algorithms are learned. Mining Text Data introduces an important niche in the text analytics field, and is an edited volume contributed by leading international researchers and practitioners focused on social networks & data mining. This book contains a wide swath in topics across social networks & data mining. Each chapter contains a comprehensive survey including the key research content on the topic, and the future directions of research in the field. There is a special focus on Text Embedded with Heterogeneous and Multimedia Data which makes the mining process much more challenging. A number of methods have been designed such as transfer learning and cross-lingual mining for such cases. Mining Text Data simplifies the content, so that advanced-level students, practitioners and researchers in computer science can benefit from this book. Academic and corporate libraries, as well as ACM, IEEE, and Management Science focused on information security, electronic commerce, databases, data mining, machine learning, and statistics are the primary buyers for this reference book Springer 2012 Cover 1 Mining Text Data 4 ISBN 9781461432227 5 Contents 6 Preface 12 Chapter 1 : AN INTRODUCTION TO TEXT MINING 14 Abstract 14 1. Introduction 14 2. Algorithms for Text Mining 17 3. Future Directions 21 References 23 Chapter 2: INFORMATION EXTRACTION FROM TEXT 24 Abstract 24 Keywords 24 1. Introduction 24 2. Named Entity Recognition 28 2.1 Rule-based Approach 29 2.2 Statistical Learning Approach 30 3. Relation Extraction 35 3.1 Feature-based Classification 36 3.2 Kernel Methods 39 3.3 Weakly Supervised Learning Methods 42 4. Unsupervised Information Extraction 43 4.1 Relation Discovery and Template Induction 44 4.2 Open Information Extraction 45 5. Evaluation 46 6. Conclusions and Summary 47 References 48 Chapter 3: A SURVEY OF TEXT SUMMARIZATION TECHNIQUES 56 Abstract 56 Keywords 57 1. How do Extractive Summarizers Work? 57 2. Topic Representation Approaches 59 2.1 Topic Words 59 2.2 Frequency-driven Approaches 61 2.3 Latent Semantic Analysis 65 2.4 Bayesian Topic Models 66 2.5 Sentence Clustering and Domain-dependent Topics 68 3. Influence of Context 69 3.1 Web Summarization 70 3.2 Summarization of Scientific Articles 71 3.3 Query-focused Summarization 71 3.4 Email Summarization 72 4. Indicator Representations and Machine Learning for Summarization 73 4.1 Graph Methods for Sentence Importance 73 4.2 Machine Learning for Summarization 75 5. Selecting Summary Sentences 77 5.1 Greedy Approaches: Maximal Marginal Relevance 77 5.2 Global Summary Selection 78 6. Conclusion 79 References 79 Chapter 4: A SURVEY OF TEXT CLUSTERING ALGORITHMS 90 Abstract 90 Keywords 90 1. Introduction 90 2. Feature Selection and Transformation Methods for Text Clustering 94 2.1 Feature Selection Methods 94 2.2 LSI-based Methods 97 2.3 Non-negative Matrix Factorization 99 3. Distance-based Clustering Algorithms 102 3.1 Agglomerative and Hierarchical Clustering Algorithms 103 3.2 Distance-based Partitioning Algorithms 105 3.3 A Hybrid Approach: The Scatter-Gather Method 107 4. Word and Phrase-based Clustering 112 4.1 Clustering with Frequent Word Patterns 113 4.2 Leveraging Word Clusters for Document Clusters 115 4.3 Co-clustering Words and Documents 116 4.4 Clustering with Frequent Phrases 118 5. Probabilistic Document Clustering and Topic Models 120 6. Online Clustering with Text Streams 123 7. Clustering Text in Networks 128 8. Semi-Supervised Clustering 131 9. Conclusions and Summary 133 References 134 Chapter 5: DIMENSIONALITY REDUCTION AND TOPIC MODELING: FROM LATENT SEMANTIC INDEXING TO LATENT DIRICHLET ALLOCATION AND BEYOND 142 Abstract 143 Keywords 143 1. Introduction 143 1.1 The Relationship Between Clustering, Dimension Reduction and Topic Modeling 144 1.2 Notation and Concepts 145 2. Latent Semantic Indexing 146 2.1 The Procedure of Latent Semantic Indexing 147 2.2 Implementation Issues 148 2.3 Analysis 150 3. Topic Models and Dimension Reduction 152 3.1 Probabilistic Latent Semantic Indexing 153 3.2 Latent Dirichlet Allocation 155 4. Interpretation and Evaluation 161 4.1 Interpretation 161 4.2 Evaluation 162 4.3 Parameter Selection 163 4.4 Dimension Reduction 163 5. Beyond Latent Dirichlet Allocation 164 5.1 Scalability 164 5.2 Dynamic Data 164 5.3 Networked Data 165 5.4 Adapting Topic Models to Applications 167 6. Conclusion 168 Acknowledgment 169 References 169 Chapter 6: A SURVEY OF TEXT CLASSIFICATION ALGORITHMS 176 Abstract 176 Keywords 176 1. Introduction 176 2. Feature Selection for Text Classification 180 2.1 Gini Index 181 2.2 Information Gain 182 2.3 Mutual Information 182 2.4 .X2-Statistic 183 2.5 Feature Transformation Methods: Supervised LSI 184 2.6 Supervised Clustering for Dimensionality Reduction 185 2.7 Linear Discriminant Analysis 186 2.8 Generalized Singular Value Decomposition 188 2.9 Interaction of Feature Selection with Classification 188 3. Decision Tree Classifiers 189 4. Rule-based Classifiers 191 5. Probabilistic and Naive Bayes Classifiers 194 5.1 Bernoulli Multivariate Model 196 5.2 Multinomial Distribution 201 5.3 Mixture Modeling for Text Classification 203 6. Linear Classifiers 206 6.1 SVM Classifiers 207 6.2 Regression-Based Classifiers 209 6.3 Neural Network Classifiers 210 6.4 Some Observations about Linear Classifiers 212 7. Proximity-based Classifiers 213 8. Classification of Linked and Web Data 216 9. Meta-Algorithms for Text Classification 222 9.1 Classifier Ensemble Learning 222 9.2 Data Centered Methods: Boosting and Bagging 223 9.3 Optimizing Specific Measures of Accuracy 224 10. Conclusions and Summary 226 References 226 Chapter 7: TRANSFER LEARNING FOR TEXT MINING 236 Abstract 236 Keywords 236 1. Introduction 237 2. Transfer Learning in Text Classification 238 2.1 Cross Domain Text Classification 238 2.2 Instance-based Transfer 244 2.3 Cross-Domain Ensemble Learning 245 2.4 Feature-based Transfer Learning for Document Classification 248 3. Heterogeneous Transfer Learning 252 3.1 Heterogeneous Feature Space 254 3.2 Heterogeneous Label Space 256 3.3 Summary 257 4. Discussion 258 5. Conclusions 259 References 260 Chapter 8: PROBABILISTIC MODELS FOR TEXT MINING 272 Abstract 272 Keywords 272 1. Introduction 273 2. Mixture Models 274 2.1 General Mixture Model Framework 275 2.2 Variations and Applications 276 2.3 The Learning Algorithms 279 3. Stochastic Processes in Bayesian Nonparametric Models 282 3.1 Chinese Restaurant Process 282 3.2 Dirichlet Process 283 3.3 Pitman-Yor Process 287 3.4 Others 288 4. Graphical Models 288 4.1 Bayesian Networks 289 4.2 Hidden Markov Models 291 4.3 Markov Random Fields 295 4.4 Conditional Random Fields 298 4.5 Other Models 299 5. Probabilistic Models with Constraints 300 6. Parallel Learning Algorithms 301 7. Conclusions 302 References 303 Chapter 9: MINING TEXT STREAMS 310 Abstract 310 Keywords 310 1. Introduction 310 2. Clustering Text Streams 312 2.1 Topic Detection and Tracking in Text Streams 320 3. Classification of Text Streams 325 4. Evolution Analysis in Text Streams 329 5. Conclusions 330 References 331 Chapter 10: TRANSLINGUAL MINING FROM TEXT DATA 336 Abstract 336 Keywords 337 1. Introduction 337 2. Traditional Translingual Text Mining – Machine Translation 338 2.1 SMT and Generative Translation Models 338 2.2 Word-Based Models 340 2.3 Phrase-Based Models 342 2.4 Syntax-Based Models 346 3. Automatic Mining of Parallel texts 349 3.1 Using Web structure 350 3.2 Matching parallel pages 352 4. Using Translation Models in CLIR 354 5. Collecting and Exploiting Comparable Texts 357 6. Selecting Parallel Sentences, Phrases and Translation Words 360 7. Mining Translingual Relations From Monolingual Texts 362 8. Mining using hyperlinks 364 9. Conclusions and Discussions 366 References 367 Chapter 11: TEXT MINING IN MULTIMEDIA 374 Abstract 374 Keywords 374 1. Introduction 375 2. Surrounding Text Mining 377 3. Tag Mining 379 3.1 Tag Ranking 379 3.2 Tag Refinement 380 3.3 Tag Information Enrichment 382 4. Joint Text and Visual Content Mining 383 4.1 Visual Re-ranking 384 5. Cross Text and Visual Content Mining 387 6. Summary and Open Issues 390 Joint text and visual content multimedia ranking 391 Scalable text mining for large-scale multimedia management 391 Multimedia social network mining 391 Acknowledgements 392 References 392 Chapter 12: TEXT ANALYTICS IN SOCIAL MEDIA 398 Abstract 398 Keywords 398 1. Introduction 398 2. Distinct Aspects of Text in Social Media 401 2.1 A General Framework for Text Analytics 401 2.2 Time Sensitivity 403 2.3 Short Length 404 2.4 Unstructured Phrases 405 2.5 Abundant Information 406 3. Applying Text Analytics to Social Media 406 3.1 Event Detection 406 3.2 Collaborative Question Answering 408 3.3 Social Tagging 410 3.4 Bridging the Semantic Gap 411 3.5 Exploiting the Power of Abundant Information 412 3.6 Related Efforts 414 4. An Illustrative Example 415 4.1 Seed Phrase Extraction 415 4.2 Semantic Feature Generation 417 4.3 Feature Space Construction 419 5. Conclusion and Future Work 420 Acknowledgments 421 References 421 Chapter 13: A SURVEY OF OPINION MINING AND SENTIMENT ANALYSIS 428 Abstract 428 Keywords 429 1. The Problem of Opinion Mining 429 1.1 Opinion Definition 429 1.2 Aspect-Based Opinion Summary 433 2. Document Sentiment Classification 435 2.1 Classification based on Supervised Learning 435 2.2 Classification based on Unsupervised Learning 437 3. Sentence Subjectivity and Sentiment Classification 439 4. Opinion Lexicon Expansion 442 4.1 Dictionary based approach 442 4.2 Corpus-based approach and sentiment consistency 443 5. Aspect-Based Sentiment Analysis 445 5.1 Aspect Sentiment Classification 446 5.2 Basic Rules of Opinions 447 5.3 Aspect Extraction 451 5.4 Simultaneous Opinion Lexicon Expansion and Aspect Extraction 453 6. Mining Comparative Opinions 454 7. Some Other Problems 457 8. Opinion Spam Detection 460 8.1 Spam Detection Based on Supervised Learning 461 8.2 Spam Detection Based on Abnormal Behaviors 462 8.3 Group Spam Detection 463 9. Utility of Reviews 464 10. Conclusions 465 References 466 Chapter 14: BIOMEDICAL TEXT MINING: A SURVEY OF RECENT PROGRESS 478 Abstract 478 Keywords 478 1. Introduction 479 2. Resources for Biomedical Text Mining 480 2.1 Corpora 480 2.2 Annotation 482 2.3 Knowledge Sources 483 2.4 Supporting Tools 484 3. Information Extraction 485 3.1 Named Entity Recognition 486 3.2 Relation Extraction 491 3.3 Event Extraction 495 4. Summarization 497 5. Question Answering 501 5.1 Medical Question Answering 502 5.2 Biological Question Answering 504 6. Literature-Based Discovery 505 7. Conclusion 508 Acknowledgements 509 References 509 Index 532 1461432227,9781461432227 Text mining applications have experienced tremendous advances because of web 2.0 and social networking applications. Recent advances in hardware and software technology have lead to a number of unique scenarios where text mining algorithms are learned. Mining Text Data introduces an important niche in the text analytics field, and is an edited volume contributed by leading international researchers and practitioners focused on social networks & data mining. This book contains a wide swath in topics across social networks & data mining. Each chapter contains a comprehensive survey including the key research content on the topic, and the future directions of research in the field. There is a special focus on Text Embedded with Heterogeneous and Multimedia Data which makes the mining process much more challenging. A number of methods have been designed such as transfer learning and cross-lingual mining for such cases. Mining Text Data simplifies the content, so that advanced-level students, practitioners and researchers in computer science can benefit from this book. Academic and corporate libraries, as well as ACM, IEEE, and Management Science focused on information security, electronic commerce, databases, data mining, machine learning, and statistics are the primary buyers for this reference book. (Quelle: buch.ch)