+612 9045 4394
 
CHECKOUT
$7.95 Delivery per order to Australia and New Zealand
100% Australian owned
Over a hundred thousand in-stock titles ready to ship
Self-Organizing Maps : Springer Series in Information Sciences - Teuvo Kohonen

Self-Organizing Maps

Springer Series in Information Sciences

Paperback Published: January 2001
ISBN: 9783540679219
Number Of Pages: 502

Share This Book:

Paperback

$259.53
or 4 easy payments of $64.88 with Learn more
Ships in 15 business days

Earn 519 Qantas Points
on this Book

Mathematical Preliminaries.- Neural Modeling.- The Basic SOM.- Physiological Interpretation of SOM.- Variants of SOM.- Learning Vector Quantization.- Applications.- Software Tools for SOM.- Hardware for SOM.- An Overview of SOM Literature.- Glossary of "Neural" Terms.- References

Mathematical Preliminariesp. 1
Mathematical Concepts and Notationsp. 2
Vector Space Conceptsp. 2
Matrix Notationsp. 8
Eigenvectors and Eigenvalues of Matricesp. 11
Further Properties of Matricesp. 13
On Matrix Differential Calculusp. 15
Distance Measures for Patternsp. 17
Measures of Similarity and Distance in Vector Spacesp. 17
Measures of Similarity and Distance Between Symbol Stringsp. 21
Averages Over Nonvectorial Variablesp. 28
Statistical Pattern Analysisp. 29
Basic Probabilistic Conceptsp. 29
Projection Methodsp. 34
Supervised Classificationp. 39
Unsupervised Classificationp. 44
The Subspace Methods of Classificationp. 46
The Basic Subspace Methodp. 46
Adaptation of a Model Subspace to Input Subspacep. 49
The Learning Subspace Method (LSM)p. 53
Vector Quantizationp. 59
Definitionsp. 59
Derivation of the VQ Algorithmp. 60
Point Density in VQp. 62
Dynamically Expanding Contextp. 64
Setting Up the Problemp. 65
Automatic Determination of Context-Independent Productionsp. 66
Conflict Bitp. 67
Construction of Memory for the Context-Dependent Productionsp. 68
The Algorithm for the Correction of New Stringsp. 68
Estimation Procedure for Unsuccessful Searchesp. 69
Practical Experimentsp. 69
Neural Modelingp. 71
Models, Paradigms, and Methodsp. 71
A History of Some Main Ideas in Neural Modelingp. 72
Issues on Artificial Intelligencep. 75
On the Complexity of Biological Nervous Systemsp. 76
What the Brain Circuits Are Notp. 78
Relation Between Biological and Artificial Neural Networksp. 79
What Functions of the Brain Are Usually Modeled?p. 81
When Do We Have to Use Neural Computing?p. 81
Transformation, Relaxation, and Decoderp. 82
Categories of ANNsp. 85
A Simple Nonlinear Dynamic Model of the Neuronp. 87
Three Phases of Development of Neural Modelsp. 89
Learning Lawsp. 91
Hebb's Lawp. 91
The Riccati-Type Learning Lawp. 92
The PCA-Type Learning Lawp. 95
Some Really Hard Problemsp. 96
Brain Mapsp. 99
The Basic SOMp. 105
A Qualitative Introduction to the SOMp. 106
The Original Incremental SOM Algorithmp. 109
The "Dot-Product SOM"p. 115
Other Preliminary Demonstrations of Topology-Preserving Mappingsp. 116
Ordering of Reference Vectors in the Input Spacep. 116
Demonstrations of Ordering of Responses in the Output Spacep. 120
Basic Mathematical Approaches to Self-Organizationp. 127
One-Dimensional Casep. 128
Constructive Proof of Ordering of Another One-Dimensional SOMp. 132
The Batch Mapp. 138
Initialization of the SOM Algorithmsp. 142
On the "Optimal" Learning-Rate Factorp. 143
Effect of the Form of the Neighborhood Functionp. 145
Does the SOM Algorithm Ensue from a Distortion Measure?p. 146
An Attempt to Optimize the SOMp. 148
Point Density of the Model Vectorsp. 152
Earlier Studiesp. 152
Numerical Check of Point Densities in a Finite One-Dimensional SOMp. 153
Practical Advice for the Construction of Good Mapsp. 159
Examples of Data Analyses Implemented by the SOMp. 161
Attribute Maps with Full Data Matrixp. 161
Case Example of Attribute Maps Based on Incomplete Data Matrices (Missing Data): "Poverty Map"p. 165
Using Gray Levels to Indicate Clusters in the SOMp. 165
Interpretation of the SOM Mappingp. 166
"Local Principal Components"p. 166
Contribution of a Variable to Cluster Structuresp. 169
Speedup of SOM Computationp. 170
Shortcut Winner Searchp. 170
Increasing the Number of Units in the SOMp. 172
Smoothingp. 175
Combination of Smoothing, Lattice Growing, and SOM Algorithmp. 176
Physiological Interpretation of SOMp. 177
Conditions for Abstract Feature Maps in the Brainp. 177
Two Different Lateral Control Mechanismsp. 178
The WTA Function, Based on Lateral Activity Controlp. 179
Lateral Control of Plasticityp. 184
Learning Equationp. 185
System Models of SOM and Their Simulationsp. 185
Recapitulation of the Features of the Physiological SOM Modelp. 188
Similarities Between the Brain Maps and Simulated Feature Mapsp. 188
Magnificationp. 189
Imperfect Mapsp. 189
Overlapping Mapsp. 189
Variants of SOMp. 191
Overview of Ideas to Modify the Basic SOMp. 191
Adaptive Tensorial Weightsp. 194
Tree-Structured SOM in Searchingp. 197
Different Definitions of the Neighborhoodp. 198
Neighborhoods in the Signal Spacep. 200
Dynamical Elements Added to the SOMp. 204
The SOM for Symbol Stringsp. 205
Initialization of the SOM for Stringsp. 205
The Batch Map for Stringsp. 206
Tie-Break Rulesp. 206
A Simple Example: The SOM of Phonemic Transcriptionsp. 207
Operator Mapsp. 207
Evolutionary-Learning SOMp. 211
Evolutionary-Learning Filtersp. 211
Self-Organization According to a Fitness Functionp. 212
Supervised SOMp. 215
The Adaptive-Subspace SOM (ASSOM)p. 216
The Problem of Invariant Featuresp. 216
Relation Between Invariant Features and Linear Subspacesp. 218
The ASSOM Algorithmp. 222
Derivation of the ASSOM Algorithm by Stochastic Approximationp. 226
ASSOM Experimentsp. 228
Feedback-Controlled Adaptive-Subspace SOM (FASSOM)p. 242
Learning Vector Quantizationp. 245
Optimal Decisionp. 245
The LVQ1p. 246
The Optimized-Learning-Rate LVQ1 (OLVQ1)p. 250
The Batch-LVQ1p. 251
The Batch-LVQ1 for Symbol Stringsp. 252
The LVQ2 (LVQ 2.1)p. 252
The LVQ3p. 253
Differences Between LVQ1, LVQ2 and LVQ3p. 254
General Considerationsp. 254
The Hypermap-Type LVQp. 256
The "LVQ-SOM"p. 261
Applicationsp. 263
Preprocessing of Optic Patternsp. 264
Blurringp. 265
Expansion in Terms of Global Featuresp. 266
Spectral Analysisp. 266
Expansion in Terms of Local Features (Wavelets)p. 267
Recapitulation of Features of Optic Patternsp. 267
Acoustic Preprocessingp. 268
Process and Machine Monitoringp. 269
Selection of Input Variables and Their Scalingp. 269
Analysis of Large Systemsp. 270
Diagnosis of Speech Voicingp. 274
Transcription of Continuous Speechp. 274
Texture Analysisp. 280
Contextual Mapsp. 281
Artifically Generated Clausesp. 283
Natural Textp. 285
Organization of Large Document Filesp. 286
Statistical Models of Documentsp. 286
Construction of Very Large WEBSOM Maps by the Projection Methodp. 292
The WEBSOM of All Electronic Patent Abstractsp. 296
Robot-Arm Controlp. 299
Simultaneous Learning of Input and Output Parametersp. 299
Another Simple Robot-Arm Controlp. 303
Telecommunicationsp. 304
Adaptive Detector for Quantized Signalsp. 304
Channel Equalization in the Adaptive QAMp. 305
Error-Tolerant Transmission of Images by a Pair of SOMsp. 306
The SOM as an Estimatorp. 308
Symmetric (Autoassociative) Mappingp. 308
Asymmetric (Heteroassociative) Mappingp. 309
Software Tools for SOMp. 311
Necessary Requirementsp. 311
Desirable Auxiliary Featuresp. 313
SOM Program Packagesp. 315
SOM_PAKp. 315
SOM Toolboxp. 317
Nenet (Neural Networks Tool)p. 318
Viscovery SOMinep. 318
Examples of the Use of SOMLPAKp. 319
File Formatsp. 319
Description of the Programs in SOM_PAKp. 322
A Typical Training Sequencep. 326
Neural-Networks Software with the SOM Optionp. 327
Hardware for SOMp. 329
An Analog Classifier Circuitp. 329
Fast Digital Classifier Circuitsp. 332
SIMD Implementation of SOMp. 337
Transputer Implementation of SOMp. 339
Systolic-Array Implementation of SOMp. 341
The COKOS Chipp. 342
The TInMANN Chipp. 342
NBISOM_25 Chipp. 344
An Overview of SOM Literaturep. 347
Books and Review Articlesp. 347
Early Works on Competitive Learningp. 348
Status of the Mathematical Analysesp. 349
Zero-Order Topology (Classical VQ) Resultsp. 349
Alternative Topological Mappingsp. 350
Alternative Architecturesp. 350
Functional Variantsp. 351
Theory of the Basic SOMp. 352
The Learning Vector Quantizationp. 358
Diverse Applications of SOMp. 358
Machine Vision and Image Analysisp. 358
Optical Character and Script Readingp. 360
Speech Analysis and Recognitionp. 360
Acoustic and Musical Studiesp. 361
Signal Processing and Radar Measurementsp. 362
Telecommunicationsp. 362
Industrial and Other Real-World Measurementsp. 362
Process Controlp. 363
Roboticsp. 364
Electronic-Circuit Designp. 364
Physicsp. 364
Chemistryp. 365
Biomedical Applications Without Image Processingp. 365
Neurophysiological Researchp. 366
Data Processing and Analysisp. 366
Linguistic and AI Problemsp. 367
Mathematical and Other Theoretical Problemsp. 368
Applications of LVQp. 369
Survey of SOM and LVQ Implementationsp. 370
Glossary of "Neural" Termsp. 373
Referencesp. 403
Indexp. 487
Table of Contents provided by Publisher. All Rights Reserved.

ISBN: 9783540679219
ISBN-10: 3540679219
Series: Springer Series in Information Sciences
Audience: General
Format: Paperback
Language: English
Number Of Pages: 502
Published: January 2001
Publisher: Springer-Verlag Berlin and Heidelberg Gmbh & Co. Kg
Country of Publication: DE
Dimensions (cm): 23.39 x 15.6  x 2.72
Weight (kg): 0.74
Edition Number: 3
Edition Type: Revised

Earn 519 Qantas Points
on this Book