Li / Cai / Kang Computational Intelligence and Intelligent Systems

4th International Symposium on Intelligence Computation and Applications, ISICA 2009, Huangshi, China, October 23-25, 2009
1. Auflage 2009
ISBN: 978-3-642-04962-0
Verlag: Springer
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, Band 51, 495 Seiten

Reihe: Communications in Computer and Information Science

Volumes CCIS 51 and LNCS 5812 constitute the proceedings of the Fourth Interational Symposium on Intelligence Computation and Applications, ISICA 2009, held in Huangshi, China, during October 23-25. ISICA 2009 attracted over 300 submissions. Through rigorous reviews, 58 papers were included in LNCS 5821,and 54 papers were collected in CCIS 51. ISICA conferences are one of the first series of international conferences on computational intelligence that combine elements of learning, adaptation, evolution and fuzzy logic to create programs as alternative solutions to artificial intelligence.

Autoren/Hrsg.

Li, Zhenhua

Cai, Zhihua

Kang, Zhuo

Inhaltsverzeichnis

1;Preface;5
2;Organization;6
3;Table of Contents;8
4;Section I: Computational Intelligence Applications;8
4.1;Omnidirectional Motion Control for the Humanoid Soccer Robot;13
4.1.1;Introduction;13
4.1.2;Preliminaries;14
4.1.2.1;Nao Model;14
4.1.2.2;Relations between Joints;14
4.1.3;Motion Control;16
4.1.3.1;Moving;16
4.1.3.2;Walking;16
4.1.3.3;Turning;17
4.1.3.4;Transverse Move;17
4.1.3.5;Symmetrical Robot Actions;17
4.1.4;Development Tool;18
4.1.4.1;Initial Module;18
4.1.4.2;Walk Module;19
4.1.5;Conclusion and Future Works;19
4.1.6;References;19
4.2;Routing Algorithm Based on Gnutella Model;21
4.2.1;Introduction;21
4.2.2;Improvement of Algorithm;22
4.2.2.1;Search Strategy of Normal Node;23
4.2.2.2;Search Strategy of Super Node;23
4.2.3;Implementation of Algorithm;25
4.2.4;The Experimental Results;26
4.2.5;Conclusion;27
4.2.6;References;27
4.3;Sliding-Window Recursive PLS Based Soft Sensing Model and Its Application to the Quality Control of Rubber Mixing Process;28
4.3.1;Introduction;28
4.3.2;Sliding-Window RPLS Model;29
4.3.2.1;The Prediction Model Based on RPLS;29
4.3.2.2;Updating Principle of Sliding-Window;30
4.3.3;Simulation Research;32
4.3.3.1;Mooney-Viscosity Prediction Based on PLS Model;32
4.3.3.2;Mooney-Viscosity Prediction Based on Sliding- Window RPLS Model;34
4.3.4;Conclusions;36
4.3.5;References;36
4.4;The Research of Method Based on Complex Multi-task Parallel Scheduling Problem;37
4.4.1;Introduction;37
4.4.2;Resource—Constrained Project Scheduling Model;38
4.4.3;Algorithm Design;38
4.4.3.1;RCPSP Problems;38
4.4.3.2;Data Structure for the Algorithm;39
4.4.3.3;The Framework of the Algorithm and the Code of Chromosome;40
4.4.3.4;Initialization Population;40
4.4.3.5;Genetic Operators Design;40
4.4.3.6;Fitness Function;41
4.4.3.7;Selection Operator;42
4.4.4;Experiments and Analysis;42
4.4.4.1;Experimental Environment;42
4.4.4.2;Experimental Data;42
4.4.4.3;The Experimental Results and Analysis;44
4.4.5;Conclusion;45
4.4.6;References;45
4.5;Towards a Bio-inspired Security Framework for Mission-Critical Wireless Sensor Networks;47
4.5.1;Introduction;47
4.5.2;Related Work;48
4.5.3;Proposed Scheme;49
4.5.3.1;Overview of MMASec;49
4.5.3.2;Model and Algorithm Selection;50
4.5.3.3;Architecture and Functions;52
4.5.3.4;Implementation Issues;53
4.5.3.5;Evaluations;53
4.5.4;Conclusion and Future Work;54
4.5.5;References;54
5;Section II: Evolutionary Algorithms;8
5.1;A Cluster-Based Orthogonal Multi-Objective Genetic Algorithm;57
5.1.1;Introduction;57
5.1.2;Problem Definition;58
5.1.3;Algorithm;58
5.1.3.1;Basic Idea;58
5.1.3.2;Algorithm Framework;59
5.1.3.3;Initialization;59
5.1.3.4;Clustering;60
5.1.3.5;Offspring Generation;62
5.1.3.6;Selection;62
5.1.4;Experiments and Results;62
5.1.4.1;Test Instances and Performance Metrics;62
5.1.4.2;Experimental Results;63
5.1.5;Conclusion;66
5.1.6;References;66
5.2;An Analysis of Semantic Aware Crossover;68
5.2.1;Introduction;68
5.2.2;Previous Work;69
5.2.3;Semantic Aware Crossover;70
5.2.4;Experimental Setup;71
5.2.5;Results and Discussion;71
5.2.5.1;Equivalent Crossovers;71
5.2.5.2;Semantic Diversity;72
5.2.5.3;Constructive Effect;74
5.2.5.4;Code Bloat;75
5.2.6;Conclusions;76
5.2.7;References;76
5.3;An Artificial Immune Univariate Marginal Distribution Algorithm;78
5.3.1;Introduction;78
5.3.2;The Univariate Marginal Distribution Algorithm;79
5.3.3;The Artificial Immune Algorithm;80
5.3.3.1;Principle of the General Artificial Immune Algorithm;80
5.3.3.2;Affinity Calculating;81
5.3.3.3;Activating and Suppressing;81
5.3.4;The Artificial Immune UMDA;82
5.3.5;Simulation Results;83
5.3.6;Conclusions;87
5.3.7;References;87
5.4;An Multi-objective Evolutionary Algorithm with Lower-Dimensional Crossover and Dense Control;88
5.4.1;Introduction;88
5.4.2;Introducing Dynamically Controlling Diversity Technique;89
5.4.3;Algorithm Description;90
5.4.3.1;Framework of Algorithm;90
5.4.3.2;Details of Algorithms;90
5.4.4;Numerical Experiments and Discussion;95
5.4.4.1;Test Problems;95
5.4.4.2;Testing Environment;95
5.4.5;Conclusions;97
5.4.6;References;98
5.5;Improved Evolutionary Programming and Its Application in Navigation System;100
5.5.1;Introduction;100
5.5.2;Improvements of Evolutionary Programming;101
5.5.2.1;Encoding Improvements-Hierarchical Evolutionary Programming;101
5.5.2.2;Mutation Improvements-Evolutionary Programming Based on Legal Individual;101
5.5.2.3;Prevention of Premature Phenomenon-Combination of Niche Technology and EP Selection;102
5.5.3;Design of BP Neural Network Based on Evolutionary Programming;103
5.5.3.1;Population Initialization;103
5.5.3.2;Learning Algorithm;103
5.5.3.3;Fitness Calculation;104
5.5.3.4;Mutation Operator;104
5.5.3.5;Select Operator;105
5.5.3.6;Hybrid Algorithm Design;105
5.5.4;Application of Improved Evolutionary Neural Network in Navigation System;105
5.5.5;Simulation Results and Conclusions;106
5.5.6;References;107
5.6;Multi-objective Emergency Facility Location Problem Based on Genetic Algorithm;109
5.6.1;Introduction;109
5.6.2;Mathematical Model of EFLP;110
5.6.3;How to Solve EFLP by GA;112
5.6.3.1;The Choice of Original Individuals;112
5.6.3.2;Fitness Function;112
5.6.3.3;GA Operators;112
5.6.3.4;Stopping Criteria;112
5.6.4;Experiments and Results;112
5.6.5;Conclusions;114
5.6.6;References;115
5.7;Trigonometric Curve Fitting Based on Genetic Algorithm and the Application of Data Processing in Geography;116
5.7.1;Introduction;116
5.7.2;The Improvement of Genetic Algorithm in the Trigonometric Functions in Curve Fitting;117
5.7.3;The Application of Trigonometric Function Fitting Based on Genetic Algorithm in Geological Data;118
5.7.4;Conclusion;120
5.7.5;References;121
6;Section III: Evolutionary Design;9
6.1;A Novel Crossover Operator in Evolutionary Algorithm for Logic Circuit Design;122
6.1.1;Introduction;122
6.1.2;Representation of Individual;123
6.1.3;Elitist Pool Evolutionary Algorithm;124
6.1.3.1;Sub-circuit Crossover Operator;124
6.1.3.2;Adaptive Mutation Strategy;125
6.1.3.3;Framework of Elitist Pool Evolutionary Algorithm;125
6.1.3.4;Evaluation;126
6.1.4;Experiment;126
6.1.5;Conclusion;128
6.1.6;References;128
6.2;Aptasensors Design Considerations;130
6.2.1;Introduction;130
6.2.2;Surface Modification;130
6.2.3;Transduction Approach;131
6.2.3.1;Electrochemical;131
6.2.3.2;Electrical;132
6.2.3.3;Optical;133
6.2.3.4;Mass;134
6.2.4;Sensor Performance;136
6.2.5;Summary and Perspective;136
6.2.6;References;137
6.3;Latent Semantic Analysis of the Languages of Life;140
6.3.1;Introduction;140
6.3.2;Mathematical Framework;141
6.3.3;Cross Language of Life;142
6.3.3.1;Empirical Results;143
6.3.4;Organism Motifs and Profiles;145
6.3.5;Phylogeny Using Doubly Singular Value Decomposition;145
6.3.6;Minimal Killer Words;147
6.3.7;Conclusion;148
6.3.8;References;148
6.4;Protein Folding Simulation by Two-Stage Optimization;150
6.4.1;Introduction;150
6.4.2;The Two-Stage Optimization;151
6.4.2.1;Sequence Conversion;152
6.4.2.2;Constraint-Based Optimal Structure Prediction in HP-Models;153
6.4.2.3;Local Search;153
6.4.3;Experiments;153
6.4.4;Conclusions;156
6.4.5;References;156
6.5;Search Direction Made Evolution Strategies Faster;158
6.5.1;Introduction;158
6.5.2;Function Optimization by Classical and Fast Evolution Strategies;159
6.5.2.1;Classical Evolution Strategies;160
6.5.2.2;Fast Evolution Strategies;160
6.5.3;Analysis the Impact of the Genetic Operators in ES;161
6.5.3.1;The Impact of Selection Operators;161
6.5.3.2;The Impact of Crossover Operators;161
6.5.3.3;The Impact of Mutation Operators;162
6.5.4;An Improved Fast Evolution Strategies;162
6.5.5;Experimental Studies;164
6.5.6;Conclusion;165
6.5.7;References;166
6.6;Topology Organization in Peer-to-Peer Platform for Genetic Algorithm Environment;168
6.6.1;Introduction;168
6.6.1.1;Related Works;168
6.6.1.2;Contribution of This Work;169
6.6.2;Design of Overlay;169
6.6.2.1;System Overview;169
6.6.2.2;Super Node Structure;170
6.6.2.3;Ordinary Node Organization;171
6.6.3;Performance Evaluations;171
6.6.3.1;Neighbor Distribution;172
6.6.3.2;SN/ON Partition;173
6.6.4;Conclusion;173
6.6.5;References;173
7;Section IV: Evolutionary Image Analysis and Signal Processing;9
7.1;Almost Periodic Solutions for Shunting Inhibitory Cellular Neural Networks with Time-Varying and Distributed Delays;174
7.1.1;Introduction;174
7.1.2;Existence of Almost Periodic Solutions;176
7.1.3;Exponential Stability of the Almost Periodic Solution;178
7.1.4;Illustrative Example;181
7.1.5;Conclusion;182
7.1.6;References;182
7.2;Applications of Computational Intelligence in Remote Sensing Image Analysis;183
7.2.1;Introduction;183
7.2.2;Neural Networks;184
7.2.3;Fuzzy Systems;184
7.2.4;Evolutionary Computation;186
7.2.5;Swarm Intelligence;187
7.2.6;Artificial Immune Systems;188
7.2.7;Conclusions;189
7.2.8;References;189
7.3;Association Relation Mining in Multimedia Data;192
7.3.1;Introduction;192
7.3.2;Association and Dependence Mining;193
7.3.3;Mining Instance of Video Strong Association and Dependence Relation;195
7.3.4;Experiment;197
7.3.5;Conclusion;198
7.3.6;References;199
7.4;DSA Image Blood Vessel Skeleton Extraction Based on Anti-concentration Diffusion and Level Set Method;200
7.4.1;Introduction;200
7.4.2;Anti-concentration Diffusion to Enhance the Blood Vessels;201
7.4.3;Otsu Local Threshold Segmentation Based on Regional Division;202
7.4.4;Vascular Skeleton Extraction Based on GMM and Fast Sweeping Method;203
7.4.5;Experiment and Analysis;206
7.4.5.1;Enhancement Effect of Vessel Using Anti-concentration Diffusion;206
7.4.5.2;Effect of Regional Division Otsu Local Threshold Segmentation;207
7.4.5.3;Effect of Skeleton Extraction;208
7.4.6;Conclusion;210
7.4.7;References;210
7.5;Space Camera Focusing Forecast Based on RBF Network;211
7.5.1;Introduction;211
7.5.2;Analysis of Network Model;212
7.5.3;Training of RBF Network;214
7.5.4;Result of Focusing Forecast Experiments;216
7.5.5;Conclusions;218
7.5.6;References;218
7.6;Wood Defect Identification Based on Artificial Neural Network;219
7.6.1;Introduction;219
7.6.2;Material and Method;219
7.6.3;Artificial Neural Network Damage Detection;220
7.6.3.1;Damage Detection Based on Wavelet Packet Energy Changes in the Artificial Neural Network;220
7.6.3.2;Based on the Frequency of Types of Structural Damage Detection Indicators;223
7.6.3.3;Damage Detection Based on the Frequency in the Neural Network;224
7.6.4;Conclusion;226
7.6.5;References;226
8;Section V: Evolutionary Optimization;10
8.1;An Improved Self-adaptive Control Parameter of Differential Evolution for Global Optimization;227
8.1.1;Introduction;227
8.1.2;Differential Evolution;228
8.1.3;Our Approach: ISADE;230
8.1.3.1;Improved Self-adaptive Control Parameter;230
8.1.3.2;Handling the Boundary Constraint of Variables;230
8.1.4;Experimental Results;231
8.1.4.1;Experimental Setup;232
8.1.4.2;Performance Criteria;232
8.1.4.3;General Performance of ISADE;233
8.1.5;Conclusion and Future Work;235
8.1.6;References;235
8.2;Distributed Evolutionary Algorithms to TSP with Ring Topology;237
8.2.1;Introductin;237
8.2.2;Distributed Evolutionary Algorithms;238
8.2.2.1;PVM;238
8.2.2.2;Algorithm Model of dEAs;238
8.2.2.3;Ring Topology;238
8.2.2.4;Master-Slave Model;238
8.2.2.5;Migration Strategy;239
8.2.2.6;Load Balancing;239
8.2.3;Experiments and Outcome Analysis;240
8.2.4;Conclusions;242
8.2.5;References;243
8.3;Self-adaptation in Fast Evolutionary Programming;244
8.3.1;Introduction;244
8.3.2;Function Optimization by Classical Evolutionary Programming;245
8.3.3;Experimental Studies;246
8.3.3.1;Benchmark Functions and Experimental Setup;246
8.3.3.2;Self-adaptation of Strategy Parameters;246
8.3.4;Conclusions;250
8.3.5;References;250
8.4;Solving SAT Problem Based on Hybrid Differential Evolution Algorithm;252
8.4.1;Introduction;252
8.4.2;SAT Problem and Its Transformation;253
8.4.2.1;SAT Problem;253
8.4.2.2;SAT Problem Is Translated into an Optimization Problem Based on {0,1};253
8.4.3;Solving SAT Problem with Hybrid Differential Evolution Algorithm;254
8.4.3.1;Chromosome Structure;254
8.4.3.2;Differential Evolution Algorithm;254
8.4.3.3;The Hill-Climbing Algorithm;256
8.4.3.4;Hybrid Differential Evolution Algorithm;256
8.4.4;Experiment;257
8.4.5;Conclusion;257
8.4.6;References;258
8.5;The Application of Evolution-Branching Algorithm on Earth-Mars Transfer Trajectory;259
8.5.1;Introduction;259
8.5.2;Evolution Programming;261
8.5.2.1;Generate an Initial Population by Orthogonal Design;261
8.5.2.2;Migration;261
8.5.2.3;Mutation;262
8.5.2.4;Mating;262
8.5.2.5;Filtering;263
8.5.3;Branching Technique;263
8.5.3.1;Branching Procedure;263
8.5.3.2;Node Branching;264
8.5.3.3;Node Evaluate;264
8.5.3.4;Node Deletion;266
8.5.4;Experiments;266
8.5.5;Conclusions;267
8.5.6;References;268
8.6;The Research of Solution to the Problems of Complex Task Scheduling Based on Self-adaptive Genetic Algorithm;269
8.6.1;Introduction;269
8.6.2;Resource-Constrained Project Scheduling Model;269
8.6.3;Using Self-adaptive Genetic Algorithm to Solve the Problems of Complex Task Scheduling;270
8.6.3.1;RCPSP Problems;270
8.6.3.2;Self-adaptive Genetic Algorithm Design;271
8.6.4;Experiments and Analysis;274
8.6.4.1;Experimental Description;274
8.6.4.2;The Experimental Results and Analysis;275
8.6.5;Conclusion;276
8.6.6;References;277
9;Section VI: Fuzzy Logic Systems;10
9.1;A Novel Approach on Designing Augmented Fuzzy Cognitive Maps Using Fuzzified Decision Trees;278
9.1.1;Introduction;278
9.1.2;Main Aspects of Fuzzy Decision Trees;279
9.1.2.1;Fuzzy Cognitive Mapping Causal Algebra;280
9.1.3;Novel Approach on Designing Augmented Fuzzy Cognitive Maps;282
9.1.4;An Illustrative Generic Example;284
9.1.5;Conclusion;286
9.1.6;References;287
9.2;Computation of Uncertain Parameters by Using Fuzzy Synthetic Decision and D-S Theory;288
9.2.1;Introduction;288
9.2.2;The Solution Method of Uncertain Parameter $p$ Based on Fuzzy Comprehensive Decision;289
9.2.2.1;The Instruction of Principle;289
9.2.2.2;An Example of Application;292
9.2.3;The Solution Method of Uncertain Parameter Based on D-S Rule;293
9.2.4;Conclusion;295
9.2.5;References;296
9.3;FAHP-Based Fuzzy Comprehensive Evaluation of M&S Credibility;297
9.3.1;Introduction;297
9.3.2;Hierarchical Evaluation Model of M&S Credibility;298
9.3.2.1;M&S Credibility;298
9.3.2.2;Hierarchical Evaluation Model of M&S Credibility;299
9.3.3;FAHP-Based FCE Model;300
9.3.3.1;Triangular Fuzzy Number and Its Operation Laws;300
9.3.3.2;FAHP-Based FCE Model;301
9.3.4;Case Study;302
9.3.5;Conclusions;304
9.3.6;References;305
9.4;Indirect Dynamic Recurrent Fuzzy Neural Network and Its Application in Identification and Control of Electro-Hydraulic Servo System;307
9.4.1;Introduction;307
9.4.2;ADRFNN;308
9.4.3;Indirect ADRFNNC;309
9.4.3.1;Design of Indirect ADRFNNC;309
9.4.3.2;Analysis on Stability of Indirect ADRFNNC;310
9.4.3.3;Modified Algorithm Preventing Parameters from Drifting;311
9.4.4;Analysis on Experiment Results;313
9.4.4.1;Experiment Results by Indirect ADRFNNC;313
9.4.4.2;Discussions on Experiment Results;314
9.4.5;Conclusion;316
9.4.6;Reference;316
9.5;Logistics Distribution Center Location Evaluation Based on Genetic Algorithm and Fuzzy Neural Network;317
9.5.1;Introduction;317
9.5.2;Fuzzy Neural Networks Structure of Logistics Distribution Center Location;318
9.5.2.1;Description of Fuzzy Neural Network;318
9.5.2.2;Network Design of Logistics Distribution Center Location;318
9.5.2.3;Model Construction;319
9.5.3;Fuzzy Neural Networks Optimized by GA;319
9.5.3.1;Basic Idea;319
9.5.3.2;Training Algorithm;320
9.5.4;Application Example;321
9.5.5;Comparative Analysis of Algorithm;323
9.5.6;Conclusion;323
9.5.7;References;323
9.6;Mean-VaR Models and Algorithms for Fuzzy Portfolio Selection;325
9.6.1;Introduction;325
9.6.2;Preliminaries;326
9.6.3;Credibilistic Mean-VaR Model;327
9.6.4;Hybrid Intelligent Algorithm for Solving Mean-VaRModel;328
9.6.5;Examples;329
9.6.6;Conclusions;330
9.6.7;References;330
9.7;MOEA-Based Fuzzy Control for Seismically Excited Structures;332
9.7.1;Introduction;332
9.7.2;Integration of Fuzzy Controller and NSGA-II;333
9.7.2.1;Definition of Control Performance Indices;333
9.7.2.2;Multi-Objective Switching Fuzzy Control Strategy;333
9.7.2.3;Multi-Objective Optimization Algorithm-NSGA-II;335
9.7.3;Numerical Simulation Analysis Using Optimization Method;336
9.7.3.1;Establishment of Seismic Loading and Structure-Damper Model;336
9.7.3.2;Optimization Result Analysis;336
9.7.4;Robustness Test with Nonlinear Numerical Simulation;339
9.7.5;Conclusions;339
9.7.6;References;340
10;Section VII: Hybrid Methods;11
10.1;A Hybrid E-Institution Model for VOs;341
10.1.1;Introduction;341
10.1.2;The Hybrid Model;342
10.1.3;Facilitation to VO;348
10.1.4;Comparison and Conclusion;353
10.1.5;References;354
10.2;An Optimal Class Association Rule Algorithm;356
10.2.1;Introduction;356
10.2.2;Basic Concept and Theory;357
10.2.3;OCARA Algorithm;358
10.2.3.1;Discovering the Optimal Rules Set;358
10.2.3.2;Sorting Rules;360
10.2.3.3;Matching Rules;360
10.2.4;Experimental Results;361
10.2.5;Conclusion;361
10.2.6;References;361
10.3;Multiple Sequence Alignment Based on Chaotic PSO;363
10.3.1;Introduction;363
10.3.2;Description of the Problem;364
10.3.2.1;Multiple Sequence Alignment (MSA);364
10.3.2.2;The Standard for Judging Multiple Sequence Alignment;365
10.3.3;Chaotic Particle Swarm Optimization;365
10.3.3.1;Particle Swarm Optimization and Its Premature to Determine;365
10.3.3.2;Chaos and Chaotic Particle Swarm Optimization;367
10.3.4;Multiple Sequence Alignment Based on Chaotic PSO;368
10.3.4.1;Relevant Definition;368
10.3.4.2;Several Problems to Be Solved;368
10.3.4.3;The Specific Steps of the Algorithm;369
10.3.5;Simulation and Results;370
10.3.6;Conclusions;371
10.3.7;References;371
10.4;Model Checking Algorithm Based on Ant Colony Swarm Intelligence;373
10.4.1;Introduction;373
10.4.1.1;Model Checking and Testing;374
10.4.1.2;Ant Colony Swarm Intelligence;375
10.4.2;Proposed Algorithm;375
10.4.2.1;Architecture and Assumptions;375
10.4.2.2;The Artificial Ants Deposit Pheromone on Traces;376
10.4.2.3;Locating Causes of the Errors According to Pheromone;377
10.4.2.4;Illustrating the Algorithm by Serving an Example;378
10.4.3;Experimental Results and Conclusions;380
10.4.4;References;380
10.5;QPSO-MD: A Quantum Behaved Particle Swarm Optimization for Consensus Pattern Identification;381
10.5.1;Introduction;381
10.5.2;DNA Motif Discovery;383
10.5.3;Problem Definition;384
10.5.4;QPSO-MD: The Proposed Framework;384
10.5.4.1;Particle Encoding;385
10.5.4.2;Fitness Function;386
10.5.4.3;Overall Dynamic of QPSO-MD;386
10.5.5;Experimental Results;387
10.5.6;Conclusion;388
10.5.7;References;389
11;Section VIII: Neural Network Architectures;11
11.1;Prediction of Hydrocarbon Reservoir Parameter Using a GA-RBF Neural Network;391
11.1.1;Introduction;391
11.1.2;GA-RBF Neural Network;392
11.1.2.1;Encode;393
11.1.2.2;Evaluation Function;393
11.1.2.3;RBF Neural Network Model Based on GA;394
11.1.3;Case Study;394
11.1.3.1;Network Input and Sample;394
11.1.3.2;Comparative Researches on Neural Network Prediction;395
11.1.4;Conclusions;397
11.1.5;References;397
11.2;Qualitative Simulation of Teachers Group Behaviors Based on BP Neural Network;399
11.2.1;Introduction;399
11.2.2;Description of Gravitation;400
11.2.2.1;Describing Gravitation Using BP Neural Network;400
11.2.2.2;Obtaining Inputs of BP Neural Network;401
11.2.2.3;The Model;403
11.2.3;Qualitative Simulation Methods;404
11.2.3.1;Variables and Their Description;404
11.2.3.2;Transition Rules;404
11.2.3.3;Filter Theory;406
11.2.3.4;Qualitative Simulation Engine;406
11.2.4;Applications;406
11.2.5;Conclusion;408
11.2.6;References;408
11.3;Stabilization of Switched Dynamic Neural Networks with Discrete Delays;410
11.3.1;Introduction;410
11.3.2;Preliminaries;411
11.3.3;Main Results;412
11.3.4;Numerical Example;415
11.3.5;Conclusions;416
11.3.6;References;417
11.4;ANN Designing Based on Co-evolutionary Genetic Algorithm with Degeneration;418
11.4.1;Introduction;418
11.4.2;Degeneration and Co-evolutionary Genetic Algorithm;419
11.4.3;Gene Coding for ANN and Fitness Function;420
11.4.4;ANN Training Based on CGA;422
11.4.5;Experiments and Results;422
11.4.6;Conclusion;424
11.4.7;References;424
11.5;Research on ACA-BP Neural Network Model;425
11.5.1;Introduction;425
11.5.2;ACA-BP Network Model;425
11.5.2.1;Basic Principle;425
11.5.2.2;Algorithm Realization;426
11.5.3;Simulation;428
11.5.4;Conclusion;430
11.5.5;References;430
12;Section IX: Predictive Modeling for Classification;11
12.1;A Text-Independent Speaker Verification System Based on Cross Entropy;431
12.1.1;Introduction;431
12.1.2;Speaker Verification System;432
12.1.2.1;Baseline System;432
12.1.2.2;Score Normalization;433
12.1.3;Approximated Cross Entropy (ACE);434
12.1.4;Experiments;436
12.1.4.1;Database;436
12.1.4.2;Evaluation Measure;436
12.1.4.3;System Description;436
12.1.4.4;Experimental Results;437
12.1.5;Conclusion;438
12.1.6;References;438
12.2;An Improved Algorithm of Apriori;439
12.2.1;Introduction;439
12.2.2;Definition of Association Rule Mining;440
12.2.2.1;Association Rule;440
12.2.2.2;$Support$ of Association Rule;440
12.2.2.3;$Confidence$ of Association Rule;440
12.2.2.4;$Strong$ Association Rule, Minimum Support Threshold, Minimum ConfidenceThreshold;440
12.2.3;Synopsis and Property of Apriori Algorithm;441
12.2.3.1;Basic Idea of Apriori Algorithm;441
12.2.3.2;Deficiency of Apriori Algorithm;441
12.2.4;Improvement of Apriori Algorithm;441
12.2.4.1;Improvement of Apriori Algorithm;441
12.2.4.2;Improved Algorithm;442
12.2.4.3;Experiments and Result;443
12.2.5;Conclusions;443
12.2.6;References;444
12.3;An Incremental Clustering with Attribute Unbalance Considered for Categorical Data;445
12.3.1;Introduction;445
12.3.2;Related Works;446
12.3.3;Our Algorithm;447
12.3.3.1;Problem Formulation;447
12.3.3.2;ICUAC Algorithm;449
12.3.4;Experiments;450
12.3.4.1;Evaluation of Clustering Result;451
12.3.4.2;Real Data Sets and Results;451
12.3.4.3;Synthetic Data Sets and Results;453
12.3.5;Conclusions;453
12.3.6;References;454
12.4;Decision Tree Classifier for Classification of Plant and Animal Micro RNA’s;455
12.4.1;Introduction;455
12.4.2;Materials and Methods;456
12.4.2.1;Software;457
12.4.2.2;Classifier;457
12.4.2.3;Evaluation;458
12.4.3;Training Set;460
12.4.4;Results and Discussion;461
12.4.5;Conclusion;462
12.4.6;References;463
12.5;Embedded Classification Learning for Feature Selection Based on K-Gravity Clustering;464
12.5.1;Introduction;464
12.5.2;Related Work;466
12.5.2.1;Feature Selection Based on Clustering;466
12.5.2.2;Feature Density Estimation;467
12.5.3;Proposed Methods;467
12.5.3.1;K-Gravity Clustering;467
12.5.3.2;Embedded Classification Learning;469
12.5.4;Experimental Results;470
12.5.5;Conclusion and Future Work;471
12.5.6;References;472
12.6;Evaluation Measures of the Classification Performance of Imbalanced Data Sets;473
12.6.1;Introduction;473
12.6.2;Commonly Performance Evaluation Measures;474
12.6.2.1;Numerical Value Performance Measure;474
12.6.2.2;Graphical Performance Analysis with Probabilistic Classifiers;475
12.6.3;Shortcomings of Some Performance Metrics;479
12.6.3.1;Shortcomings of Accuracy;479
12.6.3.2;Shortcomings of Precision/Recall;479
12.6.3.3;Shortcomings of ROC;480
12.6.4;Complex Numerical Evaluation Measures;481
12.6.4.1;F-Measure;481
12.6.4.2;G-Mean;481
12.6.4.3;Youden’s Index;482
12.6.4.4;Likelihoods;482
12.6.4.5;Discriminatory Power;482
12.6.5;Conclusions;482
12.6.6;References;483
12.7;Hybrid Classification of Pulmonary Nodules;484
12.7.1;Introduction;484
12.7.2;Existing Clustering-Based Classification Approaches;486
12.7.3;Proposed Random Forest Classification Aided by EM Clustering on Pulmonary Nodules;487
12.7.3.1;Experiment I;488
12.7.3.2;Experiment II;489
12.7.3.3;Experiment III;489
12.7.4;Discussions;490
12.7.5;Conclusion;491
12.7.6;References;491
13;Author Index;494