Detail View

Explains network dynamics, learning paradigms, and computational capabilities of feedforward, self-organization, and feedback neural network models. Specific problems such as data fusion and data modeling are addressed. Concepts and techniques are emphasized throughout, rather than their mathematical derivations. A neural network simulation software package is described and some segments of the program are given. Annotation copyright Book News, Inc. Portland, Or.

Information Provided By: :

CONTENTS
Foreword = ⅲ
Preface = ⅴ
1 Introduction = 1
 1.1 Features of Neural Networks = 2
 1.2 History of Artificial Neural Network Research = 4
 1.3 Biological Model for Artificial Neural Networks = 7
 1.4 Concepts, Building Blocks, and Terminologies = 11
 1.5 Computer Simulation of Neural Networks = 18
 1.6 Summary = 23
2 General Concepts of pattern Recognition = 35
 2.1 Non-Parametric Decision = 36
  2.1.1 Linear decision functions = 37
  2.1.2 Minimum distance classifier = 38
  2.1.3 Nearest neighbor Classifier = 39
  2.1.4 Nonlinear discriminant functions = 41
  2.1.5 Training linear classifiers = 41
 2.2 Statistical Discriminant Functions = 43
  2.2.1 Bayes machine and maximum likelihood decision = 43
  2.2.2 Normal distributed Patterns = 45
  2.2.3 Parameter estimation = 46
  2.2.4 Discriminatory measures = 48
 2.3 Cluster Seeking = 51
  2.3.1 Measures of similarity = 51
  2.3.2 A simple cluster-seeking algorithm = 53
  2.3.3 Maximum-distance algorithm = 53
  2.3.4 K-mean algorithm = 54
  2.3.5 Isodata algorithm = 55
 2.4 Conclusion = 57
3 feedforward Neural Networks = 59
 3.1 Building Blocks = 60
  3.1.1 Perceptron = 60
  3.1.2 Adaline = 63
 3.2 Multi-layered Feedforward Neural Networks = 65
  3.2.1 The capacity and classification capability = 66
  3.2.2 Learning in multi-layered neural networks = 71
  3.2.3 Variations of back-propagation algorithm = 77
 3.3 Programming Back-propagation Network = 86
 3.4 Summary = 99
4 Feedforward Neural networks for Functional Approximation = 103
 4.1 Approximation Theory and Interpolation = 105
 4.2 Principal Component Analysis Using Linear Networks = 107
  4.2.1 Learing K-L transform bases in two-layer feedforward linear network = 111
  4.2.2 Hebb learning and coordinate rotation = 114
  4.2.3 Algorithm = 116
  4.2.4 Experimental results = 118
 4.3 Neural Network Gabor Transforms = 121
  4.3.1 Gabor transform = 122
  4.3.2 Neural network for Gabor transforms = 123
 4.4 Regulization Theory and Regulization Networks = 127
  4.4.1 Regulization theory = 128
  4.4.2 Regulization Network = 131
  4.4.3 Extension of regulization networks = 132
  4.4.4 Applications = 135
 4.5 Remarks = 137
5 Applications of Feedforward Neural Networks = 141
 5.1 Adaptive Signal Processing = 142
  5.1.1 Noise canceling = 142
 5.2 Image Compression = 144
  5.2.1 Fundamental consideration for adaptive data coding = 146
  5.2.2 Composite source data model = 149
  5.2.3 Neural network adaptive image coding system = 151
  5.2.4 Block classification based on texture measures = 154
  5.2.5 Experimental results = 157
 5.3 Neural Netwok Handwritten Numeral Recognition = 160
  5.3.1 Extraction of topological salient feature points = 162
  5.3.2 Feature extraction using Fourier descriptors = 171
  5.3.3 Neural network recognition using three perspectives = 178
 5.4 Conclusion and Remarks = 186
6 Fuzzy Neural Networks = 191
 6.1 Fuzzy Set Concepts = 192
  6.1.1 Membership function = 193
  6.1.2 The geometry of fuzzy sets = 196
  6.1.3 Operations and measures on fuzzy sets = 198
  6.1.4 Measures on fuzzy sets = 198
 6.2 Fuzzy Neural Networks = 201
 6.3 Fuzzy Associative Memory = 205
  6.3.1 Fuzzification = 207
  6.3.2 Vector-matrix operations = 208
  6.3.3 Fuzzy Hebb rules = 209
  6.3.4 Reasoning with multi-input universe of discourse = 211
  6.3.5 Defuzzification = 213
  6.3.6 Learning reasoning rules by clustering = 214
  6.3.7 Network architecture for fuzzy adaptive system = 214
 6.4 Application Example Ⅰ : Backing up a Truck = 216
  6.4.1 Backing up a truck = 216
  6.4.2 Truck back up with neural network controller = 217
  6.4.3 Fuzzy back up system = 219
 6.5 Application Example Ⅱ : Financial and Economic Prediction = 224
  6.5.1 Financial data modeling and analysis = 224
  6.5.2 Stock Selection using neural networks = 226
  6.5.3 Stock selection using fuzzy networks = 228
  6.5.4 Evaluation of neural networks for stock trading = 232
 6.6 Remarks = 233
7 Competitive Learning and Self-organization = 235
 7.1 Basic Model of Competitive Learning = 237
  7.1.1 Visual pattern recognition = 240
 7.2 Interactive Activation and Competition = 245
 7.3 Kohonen Model = 249
  7.3.1 Architecture of SOM network = 250
  7.3.2 Alternative similarity measures = 252
  7.3.3 Practical hints for implementation = 252
  7.3.4 An example = 253
  7.3.5 Learning vector quantization = 254
 7.4 Applications of SOM = 256
  7.4.1 Application to speaker identification = 257
  7.4.2 Self-organization semantic map = 258
 7.5 Image Indexing Using Self-Organization Networks = 263
  7.5.1 CAFIRIIS system = 264
  7.5.2 Facial image and criminal record database system = 270
  7.5.3 Iconic index of facial images = 271
 7.6 Programming Self-Organization Networks = 286
 7.7 Conclusion and Remarks = 290
8 Adaptive Self-organization = 293
 8.1 Adapive Resonance Theory (ART) = 294
  8.1.1 ART 2 = 295
  8.1.2 Invariant visual pattern recognition with ART = 297
  8.1.3 Fuzzy ART = 299
 8.2 LEP - A Neural Network Model Based on Experiences and Perspectives = 300
  8.2.1 Motivations of LEP = 301
  8.2.2 Perspectives = 302
 8.3 A Generic Network Architecture for LEP = 305
  8.3.1 Structures of LEP neural network = 306
  8.3.2 Learning scheme = 308
  8.3.3 Fusion to get final output = 310
  8.3.4 Self-reorganization = 312
  8.3.5 Application example : image block texture classification = 314
 8.4 Supervised LEP = 316
  8.4.1 Forest inventory using remotely sensed imagery by supervised LEP = 317
  8.4.2 Ecological modeling by geographic data = 318
  8.4.3 LEP network for the forest inventory = 319
  8.4.4 The remotely sensed image data perspective = 320
  8.4.5 Fuzzy network : a perspective for the ecological model[11] = 321
  8.4.6 Forest inventory experimental results = 322
 8.5 Remarks = 324
9 Associative Memory = 327
 9.1 Basic Model = 329
  9.1.1 Pattern mathematics = 330
  9.1.2 General concepts of associative memory = 331
  9.1.3 Associative matrix = 333
  9.1.4 Association rules = 334
  9.1.5 Memory of categories = 335
 9.2 The Hopfield Model = 336
  9.2.1 Architecture of the Hopfield network = 337
  9.2.2 Learning algorithms = 339
  9.2.3 Temporal association = 343
  9.2.4 Hopfield network with analogous units = 344
  9.2.5 Hopfield network as associative memory for alphabet images = 346
 9.3 The Boltzmann Machine : Stochastic Networks = 350
  9.3.1 Boltzmann machine = 351
  9.3.2 Two-phase learning paradigm of the Boltzmann machine = 352
  9.3.3 Mean field theory learning = 354
 9.4 Bidirectional Associative Memory = 356
  9.4.1 Architecture = 356
  9.4.2 Evolution convergence = 358
  9.4.3 Storage capacity = 359
  9.4.4 Analogous BAM = 360
 9.5 Summary and Remarks = 361
10 Optimization Through Neural Networks = 365
 10.1 Optimiztion problems = 366
  10.1.1 NP-hard and NP-complete Problems = 366
  10.1.2 Optimization by constraint satisfaction = 367
  10.1.3 Identification and control of dynamic systems = 369
  10.1.4 Stability and the Lyapunov function = 370
 10.2 Optimization - Finding the Global Minima = 372
  10.2.1 Traveling salesman problem = 374
  10.2.2 Hopfield and Tank's algorithm = 374
 10.3 Image Recognition Using Neural Networks = 388
  10.3.1 The invariant features of objects = 389
  10.3.2 Recognition by hypothesis and test = 390
  10.3.3 The energy functin = 391
  10.3.4 Experimental results = 395
 10.4 Neural Network complete Boundary Extraction = 395
  10.4.1 Grid coordinates and neighborhood = 397
  10.4.2 Structural model of edges and boundaries = 398
  10.4.3 Competitive lateral interactions = 407
  10.4.4 Lateral interactions on edge grid system = 411
  10.4.5 Boundary extractin by maximum a posteriori estimation = 415
  10.4.6 Experimental results = 419
 10.5 Conclusion and Remarks = 425
Index = 429