Soft Computing in Web Information Retrieval (eBook)

Enrique Herrera-Viedma, Gabriella Pasi, Fabio Crestani (Eds.) (Herausgeber)

eBook Download: PDF

2006 | 1. Auflage
319 Seiten
Springer-Verlag
9783540315902 (ISBN)

Infos

Appl.Mathematics/Computational Methods of Engineering - Short description currently not available.

Inhalt

Preface 6
Acknowledgments 9
Foreword 10
Contents 11
Part I Document Classification 13
A Dynamic Hierarchical Fuzzy Clustering Algorithm for Information Filtering 14
1 Introduction 15
2 Categorization of Documents in IR Based 16
on Clustering Techniques 16
2.1 Partitioning Clustering Method 17
2.2 Hierarchical Clustering Method 20
divisive 20
agglomerative 20
2.3 Incremental Clustering Method 20
single-Pass 21
K-nearest neighbour 21
3 The Rationale of the Proposed Approach 21
fuzzy partition 22
fuzzy hierarchy 22
incremental dynamic clustering 22
algorithm). 22
semi-supervised fuzzy clustering 23
techniques 23
documents are represented by large and sparse vectors 23
number of clusters that must be generated at each 23
level of the hierarchy is not known 23
4 The Dynamic Fuzzy Hierarchical Clustering Algorithm 24
4.1 Documents Indexing Criteria for Clustering Purposes 24
Den 25
4.2 Input of the Clustering Algorithm 25
4.3 The Fuzzy Hierarchy of Documents 26
4.4 Generation of the Fuzzy Clusters 27
Histogram 28
Histogram( 29
4.5 Updating the Fuzzy Hierarchy with New Documents 29
5 Preliminary Experiments 30
sport 30
safety 30
game 30
math 30
lego 30
math 30
game , 30
sport 30
game ) 30
Histogram 31
lego 31
6 Conclusions 32
Acknowledgements 32
References 32
A Theoretical Framework for Web Categorization in Hierarchical Directories using Bayesian Networks 35
1 Introduction 35
2 Introduction to Bayesian Networks 37
3 Related Work on Hierarchical Categorization 38
4 Representing Hierarchical Web Directories using a Bayesian Network 39
4.1 Improving the Basic Model 43
4.2 Assessment of the Probability Distributions 44
5 Categorizing Web Pages: Inference 47
ci, 48
cj) 48
ci, cj 48
ci 48
cj 48
ci, 48
cj) 48
ci 48
cj 48
Example: 49
6 Concluding Remarks and Future Works 51
Acknowledgments 52
References 52
Personalized Knowledge Models Using RDF-Based Fuzzy Classi.cation 54
1 Introduction 54
2 Preliminary Considerations: From a Human to Machine-Oriented Vision of Information 56
2.1 Usage Scenario 58
3 Architecture Overview 58
3.1 Work.ow Scenario 61
4 Knowledge Acquisition 61
5 Features Extraction 62
5.1 Relevance Measurement of the Features 62
Collection of RDF pages: 63
Collection of schemas or dictionaries: 63
Dictionaries related to the current RDF page: 63
Accuracy: 63
Instance Relevance: 64
Property Relevance: 64
6 Rule-based Classi.cation 66
6.1 Clustering of RDF Pages 66
6.2 Rules Generation 67
If 67
then 67
7 Experimental Results 69
8 Conclusions 71
References 72
A Genetic Programming Approach for Combining Structural and Citation-Based Evidence for Text Classi.cation in Web Digital Libraries 74
1 Introduction 74
2 Background 76
3 Our Approach 77
3.1 GP System Con.gurations 77
Algorithm 1: 77
3.2 Used Terminals 78
Structural Similarity Measures 78
Citation-based Similarity Measures 79
4 The Framework for Classi.cation 81
5 Experiments 82
5.1 Sampling 83
5.2 Baselines 84
5.3 Experimental Set Up 85
5.4 Experimental Results 85
6 Related Work 87
7 Conclusion 88
Acknowledgements 88
References 89
Part II Semantic Web 93
Adding a Trust Layer to Semantic Web Metadata 94
1 Introduction 94
2 The Architecture 95
2.1 The Metadata Format 97
2.2 Modelling User Behavior in Implicit Voting 100
2.3 Trust Assertions’ Format 101
3 The Reputation Computation Problem 103
3.1 Choice of the Aggregation Operator 104
De.nition 1. 104
A1 105
A2 105
A3 105
p 105
w 105
3.2 The WOWA Operator 105
De.nition 2. 105
p 105
w 105
p 105
w 105
p 106
w, 106
a, 106
w 106
a 106
w 106
p 106
w 106
p 106
w 107
p 107
3.3 An Example 107
w 107
The Di.dent Approach 107
a 107
w 107
wn 107
p=[ 108
The Con.dent Approach 109
a 109
wn 109
4 Conclusions 109
Acknowledgments 110
References 110
A Fuzzy Linguistic Multi-agent Model Based on Semantic Web Technologies and User Pro.les 112
1 Introduction 112
2 Methodological Tools 114
2.1 Fuzzy Linguistic Tools 114
2.2 Filtering Techniques 115
2.3 Semantic Web Technologies 116
3 The Fuzzy Linguistic Multi-agent Model Based on Semantic Web and User Pro.les 116
Semantic Retrieval Phase: 117
Feedback Phase: 118
3.1 Feedback Phase: User Pro.le Updating Process 118
Step 1: 121
Step 2: 121
Step 3: 121
3.2 Feedback Phase: Recommendation Process 121
Step 1: 123
Step 2: 123
4 Example 123
5 Concluding Remarks 124
References 125
Fuzzy Concept-Based Models in Information Browsers 128
1 Introduction 128
2 Fuzziness in Concept Browsers 130
2.1 A Basic Model for Resources and Their Annotations 130
2.2 Fuzzy Elements in the Implementation 134
of Ontology-Guided Tactics 134
3 A Semantic Web Implementation of a Fuzzy Generic Concept Browser 134
4 Conclusions and Future Work 139
References 139
Evaluation of Term-based Queries using Possibilistic Ontologies 142
1 Introduction 142
2 From Fuzzy to Qualitative Pattern Matching 143
2.1 Fuzzy Pattern Matching 143
2.2 Possibilistic Ontology 145
2.3 Qualitative Pattern Matching 148
2.4 Other Approaches Using Ontologies 149
3 Using Qualitative Pattern Matching on a Database 151
Description of the 151
Platform 151
Used Ontologies 151
Examples of Queries 153
4 Retrieving Titles Using Qualitative Pattern Matching 157
4.1 Data Description 158
4.2 Examples of Queries 158
4.3 Evaluation and Results 159
5 Toward an Extension of the Approach to Full-text IR 163
5.1 Possibilistic Indexing 163
5.2 Query Evaluation 164
6 Conclusion 165
References 166
Part III Web Information Retrieval 168
Formal Theory of Connectionist Web Retrieval 169
1 Introduction 169
2 Connectionist Web Information Retrieval 171
2.1 Arti.cial Neural Network 171
Theorem 1. 172
2.2 Information Retrieval Using Multi-Layered 174
Arti.cial Neural Networks 174
2.3 Arti.cial Neural Network-based Web Retrieval – 177
A Literature Overview 177
3 Formal Theory of Connectionist Web Retrieval 180
3.1 PageRank 180
3.2 Authorities and Hubs 181
3.3 Interaction Information Retrieval 182
3.4 Interaction Information Retrieval: Particular Case 183
of the Generic Equation 183
3.5 PageRank: Particular Case of the Generic Equation 185
3.6 Hubs and Authorities: Particular Case 186
of the Generic Equation 186
4 Computational Complexity 187
in Connectionist Web Retrieval 187
4.1 Basic Concepts 187
4.2 Computational Complexity in Soft Computing-based 188
Information Retrieval – A Literature Overview 188
4.3 Computational Complexity 192
of Winner-Take-All-based Retrieval 192
Theorem 2. 192
Theorem 3. 192
Theorem 4. 192
Theorem 5. 193
5 Conclusions 195
Acknowledgements 196
References 196
Semi-fuzzy Quantifiers for Information Retrieval 201
1 Introduction 201
2 Related Work 203
3 Semi-fuzzy Quanti.ers for Information Retrieval 204
De.nition 1 (fuzzy quanti.er). 204
De.nition 2 (semi-fuzzy quanti.er). 205
De.nition 3 (quanti.er fuzzi.cation mechanism). 206
De.nition 4 ( 206
cut). 206
3.1 Query Language 208
3.2 Semantics 208
3.3 Example 209
4 Semi-fuzzy Quanti.ers and OWA Quanti.cation 210
4.1 Linguistic Quanti.cation using OWA Operators 210
4.2 Linguistic Quanti.cation using SFQ 211
4.3 Remarks 213
5 Experiments 214
5.1 Experiments: 216
5.2 Experiments: Pivoted Document Length Normalization 220
6 Conclusions and Further Work 222
Acknowledgements 223
References 223
Appendix A 225
weights 226
pivoted weights 226
Helping Users in Web Information Retrieval Via Fuzzy Association Rules 227
1 Introduction 227
2 Query Re.nement 228
3 Association Rules and Fuzzy Association Rules 229
3.1 Association Rules 230
3.2 Fuzzy Association Rules 230
3.3 Measures for Association and Fuzzy Association Rules 231
4 Query Re.nement via Fuzzy Association Rules 232
5 Document Representation for Association Rule Extraction 233
5.1 Text Transactions 234
Fuzzy Text Transactions 234
6 Extraction of Fuzzy Association Rules 235
Algorithm 1 235
6.1 The Selection of Terms for Query Re.nement 236
7 Experimental Examples 237
8 Conclusions and Future Work 240
Acknowledgements 240
References 240
Combining Soft and Hard Techniques for the Analysis of Batch Retrieval Tasks 244
1 Introduction: a Prototypical Batch Retrieval Task 244
The Statement of the Task 245
Making Some Hypotheses Explicit 246
Hypothesis 1 (Conjunctive querying). 246
Hypothesis 2 (Disjunctive querying). 247
A Decomposition of the Problem 247
Problem 1. 247
Problem 2. 248
Problem 3. 248
2 A Perfect Information Retrieval System 248
Galois Connections and Adjunctions 248
De.nition 1. 248
Proposition 1 (Polarity). 248
Proposition 2 (Axiality). 249
The Polarity of Conjunctive Querying 249
Theorem 1 (Basic theorem on Concept Lattices ([8], p. 20)). 250
Gathering Results: the Solution to Problem 1 250
3 Relevance-Induced Analysis of Retrieval Systems 251
Equivalences and Partitions De.ned by Relevance Relations 252
Rough Set Analysis of Relevance 252
An Example: the Relevance Lattice of a TREC Task 254
4 Designing the Description Mappings 256
Constraints for the Description Mappings 256
A Solution: Infomorphisms 257
5 Related Work and Discussion 258
Acknowledgements 260
References 260
Part IV Web Application 262
Search Advertising 263
1 Introduction 263
2 Basic Concepts 266
2.1 Keyword-targeted Advertising 266
2.2 Content-targeted Advertising 267
2.3 The Search Advertising Network 268
The Users 270
The Advertisers 271
The Publishers 271
3 Search Advertising Systems 272
3.1 Relevance Matching 274
dj 277
dj) 277
dj). 278
3.2 Ranking 279
3.3 Fraud Detection 281
3.4 Measurements and Feedback 282
4 Conclusions 285
References 285
Information Loss in Continuous Hybrid Microdata: Subdomain-Level Probabilistic Measures 290
1 Introduction 290
1.1 Contribution and Plan of This Paper 291
2 A Low-cost Method for Hybrid Microdata Generation 292
Algorithm 1 (Basic Procedure) 292
Algorithm 2 (Modi.cation of Matrix 293
3 Properties of the Proposed Scheme 294
3.1 Performance and Complexity 294
3.2 Data Utility 294
4 A Generic Information Loss Measure 295
5 Empirical Work 296
5.1 Information Loss and Disclosure Risk Measures 296
5.2 The Data Set 296
5.3 The Results 297
Results on the Overall Dataset 297
Top-down Generation: Posterior Subdomains 297
Bottom-up Generation: Prior Subdomains 298
6 Conclusions and Future Research 298
Acknowledgments 300
References 300
Access to a Large Dictionary of SpanishSynonyms: A Tool for Fuzzy InformationRetrieval. 302
1 Introduction 302
2 A Short Historical Introduction to Synonymy 306
3 A Computational View of Synonymy 307
4 General Architecture of an Electronic Dictionary 310
of Synonyms 310
5 Improving the Dictionary 312
6 Stand-alone Use of FDSA 314
The electronic dictionaries. 315
The algorithms that calculate the degrees of synonymy and 315
antonymy. 315
The graphical user interface. 316
7 Conclusions 317
References 318
More eBooks at www.ciando.com 0

Produkt-Details

• Erscheint lt. Verlag: 1.1.2006
• Sprache: englisch
• Themenwelt: Mathematik / Informatik ► Informatik ► Theorie / Studium
• Themenwelt: Mathematik / Informatik ► Mathematik ► Angewandte Mathematik
• Themenwelt: Technik
• ISBN-13: 9783540315902 / 9783540315902