Anmerkung: Front Cover -- Advantages and Pitfalls of Pattern Recognition -- Advantages and Pitfalls of Pattern Recognition -- Copyright -- Contents -- Preface -- Acknowledgments -- I - From data to methods -- 1 - Patterns, objects, and features -- 1.1 Objects and patterns -- 1.2 Features -- 1.2.1 Types -- 1.2.2 Feature vectors -- 1.2.3 Feature extraction -- 1.2.3.1 Delineating segments -- 1.2.3.2 Delineating regions -- 1.2.4 Transformations -- 1.2.4.1 Karhunen-Loève transformation (Principal Component Analysis) -- 1.2.4.2 Independent Component Analysis -- 1.2.4.3 Fourier transform -- 1.2.4.4 Short-time Fourier transform and spectrograms -- 1.2.4.5 Discrete wavelet transforms -- 1.2.5 Standardization, normalization, and other preprocessing steps -- 1.2.5.1 Comments -- 1.2.5.2 Outlier removal -- 1.2.5.3 Missing data -- 1.2.6 Curse of dimensionality -- 1.2.7 Feature selection -- Appendix 1 Basic notions on statistics -- A1.1 Statistical parameters of an ensemble -- A1.2 Distinction of ensembles -- 2 - Supervised learning -- 2.1 Introduction -- 2.2 Discriminant analysis -- 2.2.1 Test ban treaty-some history -- 2.2.2 The MS-mb criterion for nuclear test identification -- 2.2.3 Linear Discriminant Analysis -- 2.3 The linear perceptron -- 2.4 Solving the XOR problem: classification using multilayer perceptrons (MLPs) -- 2.4.1 Nonlinear perceptrons -- 2.5 Support vector machines (SVMs) -- 2.5.1 Linear SVM -- 2.5.2 Nonlinear SVM, kernels -- 2.6 Hidden Markov Models (HMMs)/sequential data -- 2.6.1 Background-from patterns and classes to sequences and processes -- 2.6.2 The three problems of HMMs -- 2.6.3 Including prior knowledge/model dimensions and topology -- 2.6.4 Extension to conditional random fields -- 2.7 Bayesian networks -- Appendix 2 -- Appendix 2.1 Fisher's linear discriminant analysis -- Appendix 2.2 The perceptron -- Backpropagation. , Appendix 2.3 SVM optimization of the margins -- Appendix 2.4. Hidden Markov models -- Appendix 2.4.1. Evaluation -- Appendix 2.4.2. Decoding-the Viterbi algorithm -- Appendix 2.4.3. Training-the expectation-maximization /Baum-Welch algorithm -- 3 - Unsupervised learning -- 3.1 Introduction -- 3.1.1 Metrics of (dis)similarity -- 3.1.2 Clustering -- 3.1.2.1 Partitioning clustering -- 3.1.2.1.1 Fuzzy clustering -- 3.1.2.2 Hierarchical clustering -- 3.1.2.3 Density-based clustering -- 3.2 Self-Organizing Maps -- 3.2.1 Training of an SOM -- Appendix 3 -- Appendix 3.1. Analysis of variance (ANOVA) -- Appendix 3.2 Minimum distance property for the determinant criterion -- Appendix 3.3. SOM quality -- Topological error -- Designing the map -- II - Example applications -- 4 - Applications of supervised learning -- 4.1 Introduction -- 4.2 Classification of seismic waveforms recorded on volcanoes -- 4.2.1 Signal classification of explosion quakes at Stromboli -- 4.2.2 Cross-validation issues -- 4.3 Infrasound classification -- 4.3.1 Infrasound monitoring at Mt Etna-classification with SVM -- 4.4 SVM classification of rocks -- 4.5 Inversion with MLP -- 4.5.1 Identification of parameters governing seismic waveforms -- 4.5.2 Integrated inversion of geophysical data -- 4.6 MLP in regression and interpolation -- 4.7 Regression with SVM -- 4.7.1 Background -- 4.7.2 Brief considerations on pros and cons of SVM and MLP in regression problems -- 4.8 Classification by hidden Markov models and dynamic Bayesian networks: application to seismic waveforms of tectonic, volcani ... -- 4.8.1 Background -- 4.8.2 Signals related to volcanic and tectonic activity -- 4.8.3 Classification of icequake and nonterrestrial seismic waveforms as base for further research -HMM -- 4.8.3.1 Icequakes -- 4.8.3.2 Moon quakes. , 4.8.3.3 Classification of seismic waveforms using dynamic Bayesian networks -- 4.9 Natural hazard analyses-HMMs and BNs -- 4.9.1 Estimating volcanic unrest -- 4.9.2 Reasoning under uncertainty-tsunami early warning tasks -- Appendix 4.1. Normalization issues -- Appendix 4.2. SVM Regression -- Appendix 4.3. Bias-Variance Trade-off in Curve Fitting -- 5 - Applications with unsupervised learning -- 5.1 Introduction -- 5.2 Cluster analysis of volcanic tremor data -- 5.3 Density based clustering -- 5.4 Climate zones -- 5.5 Monitoring spectral characteristics of seismic signals and volcano alert -- 5.6 Directional features -- Appendix 5 -- Appendix 5.1 Davies-Bouldin index -- Appendix 5.2 Dunn index -- Appendix 5.3 Silhouette index -- Appendix 5.4 Gap index -- Appendix 5.5 Variation of information -- III - A posteriori analysis -- 6 - A posteriori analyses-advantages and pitfalls of pattern recognition techniques -- 6.1 Introduction -- 6.2 Testing issues -- 6.3 Measuring error -- 6.4 Targets -- 6.5 Objects -- 6.6 Features and metrics -- 6.7 Concluding remarks -- 6.7.1 Multilayer perceptrons -- 6.7.2 Support Vector Machines -- 6.7.3 MLP and SVM in regression analysis -- 6.7.4 Hidden Markov models and Bayesian networks -- 6.7.5 Supervised and unsupervised learning -- 7 - Software manuals -- 7.1 Example scripts related to Chapter 2 -- 7.1.1 Linear discrimination, principal components, and marginal distributions -- 7.1.2 The perceptron -- 7.1.3 Support Vector Machines -- 7.1.4 HMM example routines (from Theodoridis et al., 2010, see http://booksite.elsevier.com/9780123744869) -- 7.2 Example scripts and programs related to Chapter 3 (unsupervised learning) -- 7.2.1 K-means clustering -- 7.2.2 Mixed models -- 7.2.3 Expectation maximization clusters -- 7.2.4 Fuzzy clustering -- 7.2.5 Hierarchical clustering -- 7.2.6 Density-based clustering. , 7.2.7 Unsupervised learning toolbox: KKAnalysis -- 7.2.7.1 Preliminaries -- 7.2.7.2 Installation -- 7.2.7.3 Files -- 7.2.7.3.1 Input files -- 7.2.7.3.2 Output files -- 7.2.7.4 Getting started -- 7.2.7.4.1 The "Input File" frame -- 7.2.7.4.2 The "figures" frame -- 7.2.7.5 Configuring KKAnalysis-the "settings" -- 7.3 Programs related to applications (Chapter 4) -- 7.3.1 Back propagation neural network (BPNN) -- 7.3.2 SVM library -- 7.4 Miscellaneous -- 7.4.1 DMGA-generating ground deformation, magnetic and gravity data -- 7.4.2 Treating fault plane solution data -- Bibliography -- Bibliography -- Index -- A -- B -- C -- D -- E -- F -- G -- H -- I -- J -- K -- L -- M -- N -- O -- P -- Q -- R -- S -- T -- U -- V -- W -- X -- Back Cover.