Focus on Artificial Neural Networks

CONTENTS......Page 6
PREFACE......Page 8
1. INTRODUCTION......Page 16
2. ARTIFICIAL NEURAL NETWORKS (ANNS)......Page 17
3. MICROEMULSIONS......Page 19
4. APPLICATION OF ANNS IN THE DEVELOPMENT OF MICROEMULSION DRUG DELIVERY SYSTEMS......Page 22
4.1. Prediction of Phase Behaviour......Page 23
4.1.1. The influence of ANNs type/architecture......Page 32
4.2. Screening of the Microemulsion Constituents......Page 33
4.3. Prediction of Structural Features of Microemulsions......Page 38
5. CONCLUSION......Page 39
REFERENCES......Page 40
ABSTRACT......Page 44
1.1. Flow Shop Scheduling......Page 45
2.1. Problem Description......Page 46
2.2.1. Initial learning stage......Page 47
2.2.2. Implementation stage......Page 48
2.4. An Illustration......Page 49
2.5. Results and Discussions......Page 51
3.1. Illustration......Page 52
3.2. Results and Discussions......Page 53
4.1. Introduction......Page 54
4.2. Architecture of the ANN......Page 55
4.3. Methodology......Page 56
4.4.2. Genetic algorithm......Page 61
4.4.3. Simulated annealing......Page 62
4.5. Results and Discussions......Page 63
4.6. Inferences......Page 65
REFERENCES......Page 66
ABSTRACT......Page 70
BRIEF DESCRIPTION OF ANN......Page 71
LITERATURE REVIEW......Page 73
ENVIRONMENTAL SCIENCES......Page 74
CHEMICAL ENGINEERING......Page 77
Modelling......Page 78
Control......Page 80
Software Sensors......Page 82
CONCLUSIONS......Page 84
REFERENCES......Page 85
ABSTRACT......Page 90
ENVIRONMENTAL FACTORS CONTROLLING WHEAT YIELD IN THE PAMPAS......Page 91
Attempts for Predicting Wheat Yield in the Pampas Using Regression Techniques......Page 92
Use of Artificial Neural Networks to Predict Wheat Yield......Page 98
Establishing Productivity Indices by an Artificial Neural Network Approach......Page 102
REFERENCES......Page 105
ABSTRACT......Page 112
1. INTRODUCTION......Page 113
2. ARTIFICIAL NEURAL NETWORKS (ANNS)......Page 115
2.1. Feedforward Neural Networks......Page 116
2.2. Recurrent Neural Networks......Page 117
2.1.1. Back Propagation neural network (BPNN)......Page 118
2.1.2. General regression neural network (GRNN)......Page 119
2.1.4. Radial basis neural network (RBNN)......Page 120
2.1.6. Self organizing map neural network (SOMNN)......Page 121
3.1.2. Influence of stepover ratio on surface roughness......Page 123
3.2. Proposed Active Suspension System (ASS)......Page 125
4. SIMULATION EXPERIMENTAL RESULTS......Page 128
NOMENCLATURE......Page 136
REFERENCES......Page 138
APPENDIX A......Page 140
APPENDIX B......Page 141
ABSTRACT......Page 144
1. INTRODUCTION......Page 145
2.1. Averaging Filter for Noise Reduction......Page 146
3.1. Architecture......Page 147
3.2. Training of a Supervised NN Filter......Page 150
4.1. Reduction of Quantum Noise in Medical X-Ray Images......Page 151
4.3. Enhancement of Contours Traced by a Cardiologist......Page 152
4.4. Separation of Ribs from Soft Tissue in Chest Radiographs......Page 154
4.5. Enhancement of Lesions in Medical Images......Page 157
4.6. Classification of Lesions and Non-Lesions in Medical Images......Page 158
ACKNOWLEDGMENTS......Page 161
REFERENCES......Page 162
ABSTRACT......Page 166
INTRODUCTION......Page 167
ARTIFICIAL NEURAL NETWORK MODELLING......Page 169
INVERSE MODEL OF DISC BRAKE PERFORMANCE......Page 172
CONCLUSION......Page 182
REFERENCES......Page 183
1. WHAT IS AN ARTIFICIAL NEURAL NETWORK (ANN)?......Page 186
2. COMMON PROPERTIES OF ANN......Page 191
3.1. New Architectures That Are Not Actually ANN......Page 193
3.2.1. Evaluation with respect to computation time......Page 198
3.2.2. Input data......Page 200
3.3. Under Evaluated Outputs......Page 201
4. CONCLUSION AND RECOMMENDATIONS......Page 202
REFERENCES......Page 203
ABSTRACT......Page 206
Tubulins and Microtubules......Page 207
Microtubules Quantum Theories......Page 209
Resonance Experiment on Microtubules......Page 210
Birefringence Experiment on Microtubules......Page 211
Materials......Page 212
Methods......Page 214
FFT Analysis of the Acquired Signals......Page 216
Statistical Analysis......Page 218
REFERENCES......Page 219
ABSTRACT......Page 224
1. INTRODUCTION......Page 225
2. METHODOLOGY......Page 226
Step 2: Determining the network structure......Page 227
2.2. Artificial Neural Network (ANN)......Page 228
2.2.2. Radial basis function network (RBFN)......Page 231
3. DATA USED......Page 234
4. ESTIMATION OF ANN PERFORMANCE EFFICIENCY......Page 235
5.3. Sensitivity Analysis of Input Parameters......Page 237
REFERENCES......Page 240
ABSTRACT......Page 246
1. INTRODUCTION......Page 247
a) Delayless filtering......Page 250
b) Signal estimation......Page 251
3. INVERTER CONTROL......Page 255
Space-vector modulation technique based on an ANN......Page 258
4. IDENTIFICATION OF MACHINE PARAMETERS......Page 260
5. NEURAL NETWORK BASED APPROACHES FOR THE EFFICIENCY IMPROVEMENT IN INDUCTION MACHINE SYSTEMS......Page 266
6. NEURAL NETWORK IMPLEMENTATIONS BY DIGITAL SIGNAL PROCESSORS AND ASIC CHIPS......Page 268
REFERENCES......Page 270
ABSTRACT......Page 272
INTRODUCTION......Page 273
Theory......Page 275
WNN Algorithm......Page 277
Training Stage......Page 278
WNNS AND THEIR APPLICATIONS TO CHEMISTRY......Page 279
WNN IN A CHEMICAL SENSING APPLICATION......Page 281
WNN Multi-Output Model......Page 282
Data Processing......Page 283
Obtained Results......Page 284
REFERENCES......Page 288
ABSTRACT......Page 292
2. SYSTEM SETUP......Page 293
3. ANN ARCHITECTURES......Page 294
3.2. Single-Layer Feedforward Network (SLFN)......Page 296
3.4. Elman Network (EN)......Page 297
3.5.2. GRNN approximator (GRNNA)......Page 298
4. PREDICTOR VALIDATION......Page 299
4.3. Block-adaptive Training......Page 300
4.4. Compression Performance......Page 302
5. EVALUATION OF ROBUSTNESS......Page 303
5.1. Ambient Noise......Page 304
5.2. Impulse Noise......Page 306
5.3. Faulty Input......Page 307
5.4. Faulty Processing Element......Page 308
6.1. Network Size......Page 309
6.4. Other Considerations......Page 310
REFERENCES......Page 311
ABSTRACT......Page 314
1. INTRODUCTION......Page 315
2. INVERSE KINEMATICS FOR SERIAL MANIPULATORS......Page 316
3. THE ARTIFICIAL NEURAL NETWORKS......Page 320
4. EXPERIMENT DESIGN (DATA COLLECTION PROCEDURE)......Page 324
5. IMPLEMENTING THE ANN......Page 326
5.1.1. The First configuration (6-6 network configuration)......Page 327
5.1.3. Networks’ performance......Page 329
5.2. Testing Phase......Page 331
REFERENCES......Page 335
ABSTRACT......Page 338
2. EXISTING FEEDFORWARD ARTIFICIAL NEURAL NETWORK MODEL......Page 339
2.2.1. Data partitioning......Page 341
2.3. Selection of Training Algorithm......Page 343
2.4. Selection of BPN Parameters......Page 345
2.4.3. Learning rate......Page 346
3. APPLICATION OF EXISTING BPN FOR CONTINUOUSLY DECREASING TIME SERIES DATA......Page 347
3.1. Existing BPN Model Results......Page 350
3.2. Modified Preprocessing Approach for Continuously Increasing and/or Decreasing Time Series Data......Page 351
4. CONCLUSION......Page 354
REFERENCES......Page 355
INTRODUCTION......Page 356
Prediction of Reaction Yield......Page 357
Prediction of Reaction Rate and Kinetic Parameters......Page 358
Prediction of Enantioselectivity......Page 359
Production of Enzyme......Page 360
Enzyme Biosensor......Page 361
Prediction of Enzyme Active Sites......Page 362
PRACTICAL CONSIDERATIONS......Page 363
REFERENCES......Page 364
ABSTRACT......Page 370
INTRODUCTION......Page 371
Neural Network Concept......Page 373
Number of Hidden Layers......Page 375
Levenberg-Marquardt Training Algorithm......Page 376
Coding in C......Page 378
Development of GUI......Page 380
Data......Page 382
Development of ANN Architecture......Page 383
Training of Networks......Page 384
Comparison of Steepest Descent and Levenberg-Marquardt Algorithms......Page 386
REFERENCES......Page 387
1. INTRODUCTION......Page 390
2.2. Formal Neuron......Page 392
3. ARTIFICIAL NEURAL NETWORKS PRINCIPLE......Page 394
3.1. Corpus Constitution......Page 396
3.3. Training Procedure......Page 397
3.3.1. Neural network parameters initialization......Page 398
3.3.2. Neuronal parameters estimation......Page 399
4. ALGORITMS OF MINIMIZATION......Page 400
4.1.1. Gradient vector of the cost......Page 401
4.1.2. Cost hessian......Page 402
4.1.3.1. Error retropropagation......Page 403
4.2.1. Gradient algorithm......Page 404
4.2.2. Newton algorithm......Page 405
4.2.3. Levenberg-marquardt algorithm......Page 406
4.3. Recursive Algorithms......Page 407
4.3.2. Gauss-newton algorithms......Page 408
6. VALIDATION PROCEDURE......Page 409
REFERENCES......Page 410
INDEX......Page 414