这个好像也没有噢……

Contents
Part I Introduction
1 Inductive Databases and Constraint-based
Data Mining: Introduction and Overview ..... 3
Saˇso Dˇzeroski
1.1 Inductive Databases ..... 3
1.2 Constraint-based Data Mining .....7
1.3 Types of Constraints ..... 9
1.4 Functions Used in Constraints .....12
1.5 KDD Scenarios ..... 14
1.6 A Brief Review of Literature Resources..... 15
1.7 The IQ (Inductive Queries for Mining Patterns and Models) Project 17
1.8 What’s in this Book .....22
2 Representing Entities in the OntoDM Data Mining Ontology.....27
Panˇce Panov, Larisa N. Soldatova, and Saˇso Dˇzeroski
2.1 Introduction ..... 27
2.2 Design Principles for the OntoDM ontology.....29
2.3 OntoDM Structure and Implementation..... 33
2.4 Identification of Data Mining Entities..... 38
2.5 Representing Data Mining Enitities in OntoDM.....46
2.6 Related Work .....52
2.7 Conclusion ..... 54
3 A Practical Comparative Study Of Data Mining Query Languages . . 59
Hendrik Blockeel, Toon Calders, ´ Elisa Fromont, Bart Goethals, Adriana
Prado, and C´eline Robardet
3.1 Introduction .....60
3.2 Data Mining Tasks..... 61
3.3 Comparison of Data Mining Query Languages.....62
3.4 Summary of the Results.....74
3.5 Conclusions .....76
xiii
xiv Contents
4 A Theory of Inductive Query Answering.....79
Luc De Raedt, Manfred Jaeger, Sau Dan Lee, and Heikki Mannila
4.1 Introduction .....80
4.2 Boolean Inductive Queries.....81
4.3 Generalized Version Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.4 Query Decomposition.....90
4.5 Normal Forms..... 98
4.6 Conclusions .....100
Part II Constraint-based Mining: Selected Techniques
5 Generalizing Itemset Mining in a Constraint Programming Setting . 107
J´er´emy Besson, Jean-Franc¸ois Boulicaut, Tias Guns, and Siegfried
Nijssen
5.1 Introduction .....107
5.2 General Concepts.....109
5.3 Specialized Approaches.....111
5.4 A Generalized Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.5 A Dedicated Solver.....116
5.6 Using Constraint Programming Systems . . . . . . . . . . . . . . . . . . . . . . 120
5.7 Conclusions .....124
6 From Local Patterns to Classification Models . . . . . . . . . . . . . . . . . . . . 127
Bj¨orn Bringmann, Siegfried Nijssen, and Albrecht Zimmermann
6.1 Introduction .....127
6.2 Preliminaries..... 131
6.3 Correlated Patterns.....132
6.4 Finding Pattern Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.5 Direct Predictions from Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
6.6 Integrated Pattern Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.7 Conclusions .....152
7 Constrained Predictive Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Jan Struyf and Saˇso Dˇzeroski
7.1 Introduction .....155
7.2 Predictive Clustering Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
7.3 Constrained Predictive Clustering Trees and Constraint Types . . . . 161
7.4 A Search Space of (Predictive) Clustering Trees . . . . . . . . . . . . . . . . 165
7.5 Algorithms for Enforcing Constraints . . . . . . . . . . . . . . . . . . . . . . . . 167
7.6 Conclusion ..... 173
8 Finding Segmentations of Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Ella Bingham
8.1 Introduction .....177
8.2 Efficient Algorithms for Segmentation . . . . . . . . . . . . . . . . . . . . . . . . 182
8.3 Dimensionality Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Contents xv
8.4 Recurrent Models.....185
8.5 Unimodal Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
8.6 Rearranging the Input Data Points . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
8.7 Aggregate Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
8.8 Evaluating the Quality of a Segmentation: Randomization . . . . . . . 191
8.9 Model Selection by BIC and Cross-validation . . . . . . . . . . . . . . . . . . 193
8.10 Bursty Sequences.....193
8.11 Conclusion ..... 194
9 Mining Constrained Cross-Graph Cliques in Dynamic Networks . . . 199
Lo¨ıc Cerf, Bao Tran Nhan Nguyen, and Jean-Franc¸ois Boulicaut
9.1 Introduction .....199
9.2 Problem Setting.....201
9.3 DATA-PEELER.....205
9.4 Extracting δ-Contiguous Closed 3-Sets . . . . . . . . . . . . . . . . . . . . . . . 208
9.5 Constraining the Enumeration to Extract 3-Cliques . . . . . . . . . . . . . 212
9.6 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
9.7 Related Work..... 224
9.8 Conclusion ..... 226
10 Probabilistic Inductive Querying Using ProbLog . . . . . . . . . . . . . . . . . 229
Luc De Raedt, Angelika Kimmig, Bernd Gutmann, Kristian Kersting,
V´ıtor Santos Costa, and Hannu Toivonen
10.1 Introduction .....229
10.2 ProbLog: Probabilistic Prolog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
10.3 Probabilistic Inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
10.4 Implementation.....238
10.5 Probabilistic Explanation Based Learning . . . . . . . . . . . . . . . . . . . . . 243
10.6 Local Pattern Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
10.7 Theory Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
10.8 Parameter Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
10.9 Application .....255
10.10 Related Work in Statistical Relational Learning . . . . . . . . . . . . . . . . 258
10.11 Conclusions .....259
Part III Inductive Databases: Integration Approaches
11 Inductive Querying with
Virtual Mining Views.....265
Hendrik Blockeel, Toon Calders, ´ Elisa Fromont, Bart Goethals, Adriana
Prado, and C´eline Robardet
11.1 Introduction .....266
11.2 The Mining Views Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
11.3 An Illustrative Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
11.4 Conclusions and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
xvi Contents
12 SINDBAD and SiQL: Overview, Applications and Future
Developments..... 289
J¨org Wicker, Lothar Richter, and Stefan Kramer
12.1 Introduction .....289
12.2 SiQL ..... 291
12.3 Example Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
12.4 A Web Service Interface for SINDBAD . . . . . . . . . . . . . . . . . . . . . . . 303
12.5 Future Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
12.6 Conclusion .....307
13 Patterns on Queries......311
Arno Siebes and Diyah Puspitaningrum
13.1 Introduction..... 311
13.2 Preliminaries.....313
13.3 Frequent Item Set Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
13.4 Transforming KRIMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
13.5 Comparing the two Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
13.6 Conclusions and Prospects for Further Research . . . . . . . . . . . . . . . 333
14 Experiment Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
Joaquin Vanschoren and Hendrik Blockeel
14.1 Introduction..... 336
14.2 Motivation .....337
14.3 Related Work.....341
14.4 A Pilot Experiment Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
14.5 Learning from the Past . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
14.6 Conclusions..... 358
Part IV Applications
15 Predicting Gene Function using Predictive Clustering Trees . . . . . . . . 365
Celine Vens, Leander Schietgat, Jan Struyf, Hendrik Blockeel, Dragi
Kocev, and Saˇso Dˇzeroski
15.1 Introduction..... 366
15.2 Related Work.....367
15.3 Predictive Clustering Tree Approaches for HMC . . . . . . . . . . . . . . . 369
15.4 Evaluation Measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
15.5 Datasets .....375
15.6 Comparison of Clus-HMC/SC/HSC . . . . . . . . . . . . . . . . . . . . . . . . . . 378
15.7 Comparison of (Ensembles of) CLUS-HMC to State-of-the-art
Methods .....380
15.8 Conclusions..... 384
Contents xvii
16 Analyzing Gene Expression Data with Predictive Clustering Trees . . 389
Ivica Slavkov and Saˇso Dˇzeroski
16.1 Introduction..... 389
16.2 Datasets .....391
16.3 Predicting Multiple Clinical Parameters . . . . . . . . . . . . . . . . . . . . . . . 392
16.4 Evaluating Gene Importance with Ensembles of PCTs . . . . . . . . . . 394
16.5 Constrained Clustering of Gene Expression Data . . . . . . . . . . . . . . . 397
16.6 Clustering gene expression time series data . . . . . . . . . . . . . . . . . . . . 400
16.7 Conclusions..... 403
17 Using a Solver Over the String Pattern Domain to Analyze Gene
Promoter Sequences.....407
Christophe Rigotti, Ieva Mitaˇsi¯unait˙e, J´er´emy Besson, Laur`ene Meyniel,
Jean-Franc¸ois Boulicaut, and Olivier Gandrillon
17.1 Introduction..... 407
17.2 A Promoter Sequence Analysis Scenario . . . . . . . . . . . . . . . . . . . . . . 409
17.3 The Marguerite Solver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
17.4 Tuning the Extraction Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
17.5 An Objective Interestingness Measure . . . . . . . . . . . . . . . . . . . . . . . . 415
17.6 Execution of the Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
17.7 Conclusion..... 422
18 Inductive Queries for a Drug Designing Robot Scientist . . . . . . . . . . . 425
Ross D. King, Amanda Schierz, Amanda Clare, Jem Rowland, Andrew
Sparkes, Siegfried Nijssen, and Jan Ramon
18.1 Introduction.....425
18.2 The Robot Scientist Eve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
18.3 Representations of Molecular Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
18.4 Selecting Compounds for a Drug Screening Library . . . . . . . . . . . . 444
18.5 Active learning.....446
18.6 Conclusions.....448
Appendix ..... 452
Author index .....455