Muutke küpsiste eelistusi

E-raamat: Advantages and Pitfalls of Pattern Recognition: Selected Cases in Geophysics

(Seismologist, Senior Researcher, Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Catania, Osservatorio Etneo, Italy), , (Seismologist, Senior Researcher, Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Catania, Osse)
  • Formaat: EPUB+DRM
  • Sari: Computational Geophysics
  • Ilmumisaeg: 23-Nov-2019
  • Kirjastus: Elsevier Science Publishing Co Inc
  • Keel: eng
  • ISBN-13: 9780128118436
Teised raamatud teemal:
  • Formaat - EPUB+DRM
  • Hind: 156,97 €*
  • * hind on lõplik, st. muud allahindlused enam ei rakendu
  • Lisa ostukorvi
  • Lisa soovinimekirja
  • See e-raamat on mõeldud ainult isiklikuks kasutamiseks. E-raamatuid ei saa tagastada.
Advantages and Pitfalls of Pattern Recognition: Selected Cases in GeophysicsSuurem pilt
  • Formaat: EPUB+DRM
  • Sari: Computational Geophysics
  • Ilmumisaeg: 23-Nov-2019
  • Kirjastus: Elsevier Science Publishing Co Inc
  • Keel: eng
  • ISBN-13: 9780128118436
Teised raamatud teemal:

DRM piirangud

  • Kopeerimine (copy/paste):

    ei ole lubatud

  • Printimine:

    ei ole lubatud

  • Kasutamine:

    Digitaalõiguste kaitse (DRM)
    Kirjastus on väljastanud selle e-raamatu krüpteeritud kujul, mis tähendab, et selle lugemiseks peate installeerima spetsiaalse tarkvara. Samuti peate looma endale  Adobe ID Rohkem infot siin. E-raamatut saab lugeda 1 kasutaja ning alla laadida kuni 6'de seadmesse (kõik autoriseeritud sama Adobe ID-ga).

    Vajalik tarkvara
    Mobiilsetes seadmetes (telefon või tahvelarvuti) lugemiseks peate installeerima selle tasuta rakenduse: PocketBook Reader (iOS / Android)

    PC või Mac seadmes lugemiseks peate installima Adobe Digital Editionsi (Seeon tasuta rakendus spetsiaalselt e-raamatute lugemiseks. Seda ei tohi segamini ajada Adober Reader'iga, mis tõenäoliselt on juba teie arvutisse installeeritud )

    Seda e-raamatut ei saa lugeda Amazon Kindle's. 

Advantages and Pitfalls of Pattern Recognition presents various methods of pattern recognition and classification, useful to geophysicists, geochemists, geologists, geographers, data analysts, and educators and students of geosciences. Scientific and technological progress has dramatically improved the knowledge of our planet with huge amounts of digital data available in various fields of Earth Sciences, such as geology, geophysics, and geography. This has led to a new perspective of data analysis, requiring specific techniques that take several features into consideration rather than single parameters. Pattern recognition techniques offer a suitable key for processing and extracting useful information from the data of multivariate analysis. This book explores both supervised and unsupervised pattern recognition techniques, while providing insight into their application.

  • Offers real-world examples of techniques for pattern recognition and handling multivariate data
  • Includes examples, applications, and diagrams to enhance understanding
  • Provides an introduction and access to relevant software packages
Preface xi
Acknowledgments xvii
Part I From data to methods
Chapter 1 Patterns, objects, and features
3(30)
1.1 Objects and patterns
3(1)
1.2 Features
3(24)
1.2.1 Types
3(1)
1.2.2 Feature vectors
4(1)
1.2.3 Feature extraction
5(1)
1.2.3.1 Delineating segments
5(3)
1.2.3.2 Delineating regions
8(2)
1.2.4 Transformations
10(1)
1.2.4.1 Karhunen---Loeve transformation (Principal Component Analysis)...
10(2)
1.2.4.2 Independent Component Analysis
12(1)
1.2.4.3 Fourier transform
13(1)
1.2.4.4 Short-time Fourier transform and spectrograms
14(1)
1.2.4.5 Discrete wavelet transforms
15(6)
1.2.5 Standardization, normalization, and other preprocessing steps
21(1)
1.2.5.1 Comments
21(1)
1.2.5.2 Outlier removal
22(1)
1.2.5.3 Missing data
22(1)
1.2.6 Curse of dimensionality
23(1)
1.2.7 Feature selection
24(3)
Appendix 1 Basic notions on statistics
27(6)
A1.1 Statistical parameters of an ensemble
27(4)
A1.2 Distinction of ensembles
31(2)
Chapter 2 Supervised learning
33(54)
2.1 Introduction
33(2)
2.2 Discriminant analysis
35(5)
2.2.1 Test ban treaty---some history
35(1)
2.2.2 The Ms---mb criterion for nuclear test identification
35(1)
2.2.3 Linear Discriminant Analysis
36(4)
2.3 The linear perceptron
40(6)
2.4 Solving the XOR problem: classification using multilayer perceptrons (MLPs)
46(7)
2.4.1 Nonlinear perceptrons
48(5)
2.5 Support vector machines (SVMs)
53(6)
2.5.1 Linear SVM
53(3)
2.5.2 Nonlinear SVM, kernels
56(3)
2.6 Hidden Markov Models (HMMs)/sequential data
59(10)
2.6.1 Background---from patterns and classes to sequences and processes
59(3)
2.6.2 The three problems of HMMs
62(5)
2.6.3 Including prior knowledge/model dimensions and topology
67(1)
2.6.4 Extension to conditional random fields
68(1)
2.7 Bayesian networks
69(4)
Appendix 2
73(14)
Appendix 2.1 Fisher's linear discriminant analysis
73(3)
Appendix 2.2 The perceptron
76(2)
Appendix 2.3 SVM optimization of the margins
78(2)
Appendix 2.4 Hidden Markov models
80(1)
Appendix 2.4.1 Evaluation
80(2)
Appendix 2.4.2 Decoding---the Viterbi algorithm
82(1)
Appendix 2.4.3 Training---the expectation---maximization /Baum---Welch algorithm
83(4)
Chapter 3 Unsupervised learning
87(40)
3.1 Introduction
87(27)
3.1.1 Metrics of (dis)similarity
88(2)
3.1.2 Clustering
90(1)
3.1.2.1 Partitioning clustering
91(13)
3.1.2.2 Hierarchical clustering
104(5)
3.1.2.3 Density-based clustering
109(5)
3.2 Self-Organizing Maps
114(5)
3.2.1 Training of an SOM
117(2)
Appendix 3
119(8)
Appendix 3.1 Analysis of variance (ANOVA)
119(1)
Appendix 3.2 Minimum distance property for the determinant criterion
120(1)
Appendix 3.3 SOM quality
121(6)
Part II Example applications
Chapter 4 Applications of supervised learning
127(62)
4.1 Introduction
127(1)
4.2 Classification of seismic waveforms recorded on volcanoes
128(11)
4.2.1 Signal classification of explosion quakes at Stromboli
132(5)
4.2.2 Cross-validation issues
137(2)
4.3 Infrasound classification
139(5)
4.3.1 Infrasound monitoring at Mt Etna---classification with SVM
141(3)
4.4 SVM classification of rocks
144(7)
4.5 Inversion with MLP
151(7)
4.5.1 Identification of parameters governing seismic waveforms
151(1)
4.5.2 Integrated inversion of geophysical data
152(6)
4.6 MLP in regression and interpolation
158(5)
4.7 Regression with SVM
163(5)
4.7.1 Background
163(4)
4.7.2 Brief considerations on pros and cons of SVM and MLP in regression problems
167(1)
4.8 Classification by hidden Markov models and dynamic Bayesian networks: application to seismic waveforms of tectonic, volcanic and lunar origin
168(12)
4.8.1 Background
168(1)
4.8.2 Signals related to volcanic and tectonic activity
169(5)
4.8.3 Classification of icequake and nonterrestrial seismic waveforms as base for further research ---HMM
174(1)
4.8.3.1 Icequakes
174(2)
4.8.3.2 Moon quakes
176(3)
4.8.3.3 Classification of seismic waveforms using dynamic Bayesian networks
179(1)
4.9 Natural hazard analyses---HMMs and BNs
180(3)
4.9.1 Estimating volcanic unrest
180(1)
4.9.2 Reasoning under uncertainty---tsunami early warning tasks
181(2)
Appendix 4.1 Normalization issues
183(1)
Appendix 4.2 SVM Regression
184(1)
Appendix 4.3 Bias---Variance Trade-off in Curve Fitting
185(4)
Chapter 5 Applications with unsupervised learning
189(48)
5.1 Introduction
189(2)
5.2 Cluster analysis of volcanic tremor data
191(6)
5.3 Density based clustering
197(6)
5.4 Climate zones
203(10)
5.5 Monitoring spectral characteristics of seismic signals and volcano alert
213(9)
5.6 Directional features
222(8)
Appendix 5
230(7)
Appendix 5.1 Davies-Bouldin index
230(1)
Appendix 5.2 Dunn index
231(1)
Appendix 5.3 Silhouette index
232(1)
Appendix 5.4 Gap index
232(1)
Appendix 5.5 Variation of information
233(4)
Part III A posteriori analysis
Chapter 6 A posteriori analyses---advantages and pitfalls of pattern recognition techniques
237(24)
6.1 Introduction
237(1)
6.2 Testing issues
238(1)
6.3 Measuring error
239(7)
6.4 Targets
246(5)
6.5 Objects
251(2)
6.6 Features and metrics
253(4)
6.7 Concluding remarks
257(4)
6.7.1 Multilayer perceptrons
257(1)
6.7.2 Support Vector Machines
258(1)
6.7.3 MLP and SVM in regression analysis
258(1)
6.7.4 Hidden Markov models and Bayesian networks
258(1)
6.7.5 Supervised and unsupervised learning
259(2)
Chapter 7 Software manuals
261(54)
7.1 Example scripts related to
Chapter 2
261(14)
7.1.1 Linear discrimination, principal components, and marginal distributions
261(6)
7.1.2 The perceptron
267(1)
7.1.3 Support Vector Machines
268(4)
7.1.4 HMM example routines
272(3)
7.2 Example scripts and programs related to
Chapter 3 (unsupervised learning)
275(25)
7.2.1 K-means clustering
275(1)
7.2.2 Mixed models
275(3)
7.2.3 Expectation maximization clusters
278(1)
7.2.4 Fuzzy clustering
279(1)
7.2.5 Hierarchical clustering
280(1)
7.2.6 Density-based clustering
280(2)
7.2.7 Unsupervised learning toolbox: KKAnalysis
282(1)
7.2.7.1 Preliminaries
282(1)
7.2.7.2 Installation
283(1)
7.2.7.3 Files
283(2)
7.2.7.4 Getting started
285(11)
7.2.7.5 Configuring KKAnalysis---the "settings"
296(4)
7.3 Programs related to applications (Chapter 4)
300(7)
7.3.1 Back propagation neural network (BPNN)
300(2)
7.3.2 SVM library
302(5)
7.4 Miscellaneous
307(8)
7.4.1 DMGA---generating ground deformation, magnetic and gravity data
307(4)
7.4.2 Treating fault plane solution data
311(4)
Bibliography 315(12)
Index 327
Horst Langer has developed methods for automatic alert systems and early warning on Mount Etna as well as tools that are routinely operated in the monitoring room of the institute and are part of the alert system for Civil Protection. Aside from his documented experience in the application of various pattern recognition techniques, he has also published computer programs for pattern recognition. Susanna Falsaperla has a long experience in the application of pattern recognition techniques and was among the first seismologists to apply automatic classification to seismic signals on volcanoes. She has made extensive use of pattern recognition in volcanology to relate multidisciplinary data to volcanic unrest and eruptive activity. Conny Hammer has worked on automatic classification of seismic signals in continuous data streams and has introduced novel concepts and tools into the seismological community from fields of machine learning (e.g., speech processing). Her automatic recognition tools are currently implemented in daily observatory routines. Besides automatic event detection, she has focused on the application of machine learning tools in seismic site characterization.
Lisainfo e-raamatute kohta Lisainfo e-raamatute kohta
Püsilink: https://www.kriso.ee/db/97801281184366e.html
Märksõnad: