An Introduction to Machine Learning - Miroslav Kubat

An Introduction to Machine Learning

(Autor)

Buch | Hardcover
XIII, 348 Seiten
2017 | 2nd ed. 2017
Springer International Publishing (Verlag)
978-3-319-63912-3 (ISBN)
74,89 inkl. MwSt
zur Neuauflage
  • Titel erscheint in neuer Auflage
  • Artikel merken
Zu diesem Artikel existiert eine Nachauflage
lt;p>This textbook presents fundamental machine learning concepts in an easy to understand manner by providing practical advice, using straightforward examples, and offering engaging discussions of relevant applications. The main topics include Bayesian classifiers, nearest-neighbor classifiers, linear and polynomial classifiers, decision trees, neural networks, and support vector machines. Later chapters show how to combine these simple tools by way of "boosting," how to exploit them in more complicated domains, and how to deal with diverse advanced practical issues. One chapter is dedicated to the popular genetic algorithms.

This revised edition contains three entirely new chapters on critical topics regarding the pragmatic application of machine learning in industry. The chapters examine multi-label domains, unsupervised learning and its use in deep learning, and logical approaches to induction. Numerous chapters have been expanded, and the presentation of the material has been enhanced. The book contains many new exercises, numerous solved examples, thought-provoking experiments, and computer assignments for independent work.

Miroslav Kubat, Associate Professor at the University of Miami, has been teaching and studying machine learning for over 25 years. He has published more than 100 peer-reviewed papers, co-edited two books, served on the program committees of over 60 conferences and workshops, and is an editorial board member of three scientific journals. He is widely credited with co-pioneering research in two major branches of the discipline: induction of time-varying concepts and learning from imbalanced training sets. He also contributed to research in induction from multi-label examples, induction of hierarchically organized classes, genetic algorithms, and initialization of neural networks.

lt;b>1              A Simple Machine-Learning Task                                                               1

1.1         Training  Sets and Classifiers.......................................................................... 1

1.2         Minor Digression:  Hill-Climbing Search....................................................... 5

1.3         Hill Climbing in  Machine Learning................................................................ 9

1.4         The Induced Classifier's Performance........................................................ 12

1.5         Some Di culties with  Available Data......................................................... 14

1.6         Summary and Historical Remarks............................................................... 18

1.7         Solidify Your Knowledge.............................................................................. 19

2              Probabilities:  Bayesian Classifiers                                                                22

2.1         The Single-Attribute Case............................................................................. 22

2.2         Vectors  of Discrete Attributes..................................................................... 27

2.3         Probabilities of Rare Events:  Exploiting the   Expert's Intuition............. 29

2.4         How  to Handle Continuous Attributes....................................................... 35

2.5         Gaussian "Bell" Function:  A  Standard pdf................................................. 38

2.6         Approximating PDFs with Sets  of Gaussians............................................ 40

2.7         Summary and Historical Remarks............................................................... 43

2.8         Solidify Your Knowledge.............................................................................. 46

3              Similarities:  Nearest-Neighbor Classifiers                                                 49

3.1         The k-Nearest-Neighbor Rule...................................................................... 49

3.2         Measuring Similarity...................................................................................... 52

3.3         Irrelevant  Attributes and Scaling Problems............................................... 56

3.4         Performance Considerations........................................................................ 60

3.5         Weighted Nearest Neighbors....................................................................... 63

3.6         Removing Dangerous Examples.................................................................. 65

3.7         Removing Redundant Examples.................................................................. 68

3.8         Summary and Historical Remarks............................................................... 71

3.9         Solidify Your Knowledge.............................................................................. 72


 

 

 

 

4              Inter-Class Boundaries:

Linear and Polynomial Classifiers                                                                  75

4.1         The Essence..................................................................................................... 75

4.2         The Additive Rule:  Perceptron Learning.................................................... 79

4.3         The  Multiplicative  Rule:  WINNOW............................................................ 85

4.4         Domains with More than  Two Classes........................................................ 88

4.5         Polynomial Classifiers..................................................................................... 91

4.6         Specific Aspects of Polynomial Classifiers................................................... 93

4.7         Numerical Domains and Support Vector Machines................................... 97

4.8         Summary and Historical Remarks.............................................................. 100

4.9         Solidify Your Knowledge............................................................................. 101

5              Artificial Neural Networks                                                                            105

5.1         Multilayer Perceptrons as Classifiers.......................................................... 105

5.2         Neural Network's Error............................................................................... 110

5.3         Backpropagation of Error........................................................................... 111

5.4         Special Aspects of Multilayer Perceptrons................................................ 117

5.5         Architectural Issues...................................................................................... 121

5.6         Radial Basis Function Networks................................................................. 123

5.7         Summary and Historical Remarks.............................................................. 126

5.8         Solidify Your Knowledge............................................................................. 128

6              Decision  Trees                                                                                                    130

6.1         Decision Trees

6.2         Induction of Decision Trees........................................................................ 134

6.3         How Much Information Does an   Attribute Convey?............................... 137

6.4         Binary Split of a   Numeric Attribute.......................................................... 142

6.5         Pruning.......................................................................................................... 144

6.6         Converting the Decision Tree  into Rules.................................................. 149

6.7         Summary and Historical Remarks.............................................................. 151

6.8         Solidify Your Knowledge............................................................................. 153

7              Computational Learning Theory                                                                  157

7.1         PAC Learning................................................................................................. 157

7.2         Examples  of PAC  Learnability.................................................................... 161

7.3         Some Practical and Theoretical Consequences......................................... 164

7.4         VC-Dimension and Learnability................................................................. 166

7.5         Summary and Historical Remarks.............................................................. 169

7.6         Exercises and Thought Experiments......................................................... 170


 

 

 

 

8              A  Few  Instructive Applications                                                                   173

8.1         Character Recognition................................................................................ 173

8.2         Oil-Spill Recognition.................................................................................... 177

8.3         Sleep Classification...................................................................................... 181

8.4         Brain-Computer Interface.......................................................................... 185

8.5         Medical Diagnosis........................................................................................ 189

8.6         Text Classification........................................................................................ 192

8.7         Summary and Historical Remarks............................................................ 194

8.8         Exercises and Thought Experiments........................................................ 195

9              Induction  of Voting Assemblies                                                                  198

9.1         Bagging.......................................................................................................... 198

9.2         Schapire's Boosting..................................................................................... 201

9.3         Adaboost:  Practical Version of Boosting................................................. <205

9.4         Variations on the  Boosting Theme........................................................... 210

9.5         Cost-Saving Benefits of  the Approach...................................................... 213

9.6         Summary and Historical Remarks............................................................ 215

9.7         Solidify Your Knowledge............................................................................ 216

10     Some  Practical  Aspects  to Know About                                                   219

10.1     A Learner's Bias.......................................................................................... 219

10.2     Imbalanced Training Sets........................................................................... 223

10.3     Context-Dependent Domains..................................................................... 228

10.4     Unknown Attribute Values......................................................................... 231

10.5     Attribute Selection....................................................................................... 234

10.6     Miscellaneous............................................................................................... 237

10.7     Summary and Historical Remarks............................................................ 238

10.8     Solidify Your Knowledge............................................................................ 240

11     Performance Evaluation                                                                                 243

11.1     Basic Performance Criteria........................................................................ 243

11.2     Precision and Recall.................................................................................... 247

11.3     Other Ways  to Measure Performance..................................................... 252

11.4     Learning Curves and  Computational Costs............................................. 255

11.5     Methodologies of Experimental Evaluation............................................. 258

11.6     Summary and Historical Remarks............................................................ 261

11.7     Solidify Your Knowledge............................................................................ 263


 

 

 

 

12     Statistical Significance                                                                                     266

12.1     Sampling a Population................................................................................ 266

12.2     Benefiting from the  Normal Distribution................................................ 271

12.3     Confidence Intervals................................................................................... 275

12.4     Statistical Evaluation of  a Classifier.......................................................... 277

12.5     Another Kind of  Statistical Evaluation..................................................... 280

12.6     Comparing Machine-Learning Techniques.............................................. 281

12.7     Summary and Historical Remarks............................................................ 284

12.8     Solidify Your Knowledge............................................................................ 285<

13     Induction  in Multi-Label Domains                                                              287

13.1     Classical Machine Learning in

Multi-Label Domains................................................................................... 287

13.2     Treating  Each  Class Separately:

Binary Relevance......................................................................................... 290

13.3     Classifier Chains........................................................................................... 293

13.4     Another Possibility: Stacking..................................................................... 296

13.5     A Note on Hierarchically  Ordered Classes............................................... 298

13.6     Aggregating the Classes.............................................................................. 301

13.7     Criteria for Performance Evaluation........................................................ 304

13.8     Summary and Historical Remarks............................................................ 307

13.9     Solidify Your Knowledge............................................................................ 308

14     Unsupervised Learning                                                                                    311

14.1     Cluster Analysis........................................................................................... 311

14.2     A Simple Algorithm: k-Means.................................................................... 315

14.3     More Advanced Versions  of k-Means...................................................... 321

14.4     Hierarchical Aggregation............................................................................ 323

14.5     Self-Organizing Feature Maps: Introduction........................................... 326

14.6     Some Important Details.............................................................................. 329

14.7     Why Feature Maps?.................................................................................... 332

14.8     Summary and Historical Remarks............................................................ 334

14.9     Solidify Your Knowledge............................................................................ 335

15     Classifiers in the Form   of Rulesets                                                           338

15.1     A Class Described  By Rules....................................................................... 338

15.2     Inducing Rulesets by  Sequential Covering............................................... 341

15.3     Predicates and Recursion.......................................................................... 344

15.4     More Advanced Search Operators............................................................ 347


 

 

 

 

15.5     Summary and Historical Remarks.............................................................. 349

15.6     Solidify Your Knowledge............................................................................ 350

16     The Genetic Algorithm<                                                                                    352<

16.1     The Baseline Genetic Algorithm................................................................ 352

16.2     Implementing the Individual Modules...................................................... 355

16.3     Why it Works............................................................................................... 359

16.4     The Danger of  Premature Degeneration................................................. 362

16.5     Other Genetic Operators............................................................................ 364

16.6     Some Advanced Versions........................................................................... 367

16.7     Selections in k-NN Classifiers..................................................................... 370

16.8     Summary and Historical Remarks............................................................ 373

16.9     Solidify Your Knowledge............................................................................ 374

17     Reinforcement Learning                                                                                 376

17.1     How  to Choose the Most  Rewarding Action........................................... 376

17.2     States and Actions in  a Game.................................................................... 379

17.3     The SARSA Approach................................................................................. 383

17.4     Summary and Historical Remarks............................................................ 384

17.5     Solidify Your Knowledge............................................................................ 384

Index                                                                                                                           395

"The presentation is mainly empirical, but precise and pedagogical, as each concept introduced is followed by a set of questions which allows the reader to check immediately whether they understand the topic. Each chapter ends with a historical summary and a series of computer assignments. ... this book could serve as textbook for an undergraduate introductory course on machine learning ... ." (Gilles Teyssière, Mathematical Reviews, April, 2017)

"This book describes ongoing human-computer interaction (HCI) research and practical applications. ... These techniques can be very useful in AR/VR development projects, and some of these chapters can be used as examples and guides for future research." (Miguel A. Garcia-Ruiz, Computing Reviews, January, 2019)

Erscheinungsdatum
Zusatzinfo XIII, 348 p. 85 illus., 3 illus. in color.
Verlagsort Cham
Sprache englisch
Maße 155 x 235 mm
Gewicht 707 g
Themenwelt Informatik Datenbanken Data Warehouse / Data Mining
Informatik Theorie / Studium Künstliche Intelligenz / Robotik
Mathematik / Informatik Mathematik Finanz- / Wirtschaftsmathematik
Schlagworte Artificial Intelligence • artificial intelligence (incl. robotics) • bayesian classifiers • Big Data/Analytics • Boosting • business mathematics & systems • Business mathematics & systems • Computational Intelligence • Computational Learning Theory • Computer Science • Data Mining • data mining and knowledge discovery • decision trees • Deep learning • Expert systems / knowledge-based systems • Genetic algorithms • linear and polynomial classifiers • machine learning • nearest neighbor classifier • Neural networks • Performance Evaluation • Reinforcement Learning • Robotics • Statistical Learning • time-varying classes, imbalanced representation • Unsupervised Learning
ISBN-10 3-319-63912-9 / 3319639129
ISBN-13 978-3-319-63912-3 / 9783319639123
Zustand Neuware
Haben Sie eine Frage zum Produkt?
Mehr entdecken
aus dem Bereich