Fuzzy Logic and Neural Networks for Hybrid Intelligent System Design

Description

This book covers recent developments on fuzzy logic, neural networks and optimization algorithms, as well as their hybrid combinations. In addition, the above-mentioned methods are applied to areas such as intelligent control and robotics, pattern recognition, medical diagnosis, time series prediction and optimization of complex problems. Nowadays, the main topic of the book is highly relevant, as most current intelligent systems and devices in use utilize some form of intelligent feature to enhance their performance. In addition, on the theoretical side, new and advanced models and algorithms of type-2 and type-3 fuzzy logic are presented, which are of great interest to researchers working on these areas. Also, new nature-inspired optimization algorithms and innovative neural models are put forward in the manuscript, which are very popular subjects, at this moment. There are contributions on theoretical aspects as well as applications, which make the book very appealing to a wide audience, ranging from researchers to professors and graduate students working in the theory and applications of the computational intelligence area. Inspired by the flight behavior and mating process of mayflies, the Mayfly algorithm combines the main advantages of swarm intelligence and evolutionary algorithms, resulting in better performance than the particle swarm algorithm. So, we proposed a modification of Mayfly algorithm by applying a fuzzy parameter adapter to be able to apply this to neural network problems. We were able to observe that the fuzzy adapter improves the speed of convergence of the mayfly algorithm and when applied to a neural network for the Mackey glass series, it manages to detect the optimal number of neurons of the hidden layer for the network architecture. However, when using the Mayfly algorithm to optimize the architecture of neural networks, the results do not improve much, so we can deduce that this metaheuristic method is not recommended (for the moment) for this type of optimization, due to the fact that the root mean square error did not get below 0.001 even using the modified Mayfly algorithm with the fuzzy adapter. There are a total of 14 papers forming the book in the above-mentioned topics. In conclusion, the edited book comprises papers on diverse aspects of fuzzy logic, neural networks and nature-inspired optimization meta-heuristics for designing and implementing hybrid intelligent systems and their application in areas , such as intelligent control and robotics, pattern recognition, time series prediction and optimization of complex problems. Preface About This Book Contents On Decision Making Applications via Distance Measures 1 Introduction 2 Preliminaries 3 An Application of Intuitionistic Fuzzy Logic with Distance Measure 3.1 Comparison of Distance Measures with the Application of Intuitionistic Fuzzy Sets 3.2 Determination of the High School Where the Students Will Enroll Using the Normalized Hamming Distance Measure 4 An Application of Intuitionistic Fuzzy Logic with Similarity Measure in Decision Making 5 Conclusion References On Intuitionistic Fuzzy Abstract Algebras 1 Introduction 2 Preliminaries 3 Intuitionistic Fuzzy Abstract Algebras 3.1 Intuitionistic Fuzzy Congruence Relations on Abstract Algebras 3.2 Isomorphism Theorems on Intuitionistic Fuzzy Universal Algebras 4 Conclusion References Generalization of Intuitionistic Fuzzy Submodules of a Module by Using Triangular Norms and Conorms and (T,S)-L Subrings 1 Introduction 2 Preliminaries 3 Intuitionistic (T,S)-Fuzzy Submodules 4 Intuitionistic (T, S)-L Fuzzy Subrings References Fuzzy Dynamic Parameter Adaptation in the Mayfly Algorithm: Implementation of Fuzzy Adaptation and Tests on Benchmark Functions and Neural Networks 1 Introduction 2 Theoretical Framework 3 Proposed Method 4 Study of the Impact of the Parameters 5 Fuzzy Adapter Design 5.1 Fuzzy Adapter with 1 Input and 1 Output 5.2 Fuzzy Adapter with 2 Inputs and 2 Outputs 6 Mackey Glass Neural Network Design 7 Results 8 Conclusions References Fuzzy Classifier Using the Particle Swarm Optimization Algorithm for the Diagnosis of Arterial Hypertension 1 Introduction 2 Proposed Type-1 and Type-2 Fuzzy Classifier 2.1 Blood Pressure 2.2 Rules of Fuzzy Systems 2.3 Particle Swarm Optimization for the Fuzzy System 3 Experimental Results 4 Conclusion References A Survey of Models and Solution Methods for the Internet Shopping Optimization Problem 1 Introduction 2 Literature Review 2.1 Internet Shopping Optimization Problem with Shipping Costs 2.2 Internet Shopping Optimization Problem with Shipping Costs and Discounts 2.3 Internet Shopping Optimization Problem with a Budget (B-ISOP) 2.4 Internet Shopping Optimization Problem with Price Sensitive Discounts 2.5 Trusted Internet Shopping Optimization Problem (T-ISOP) 2.6 Bi-objective Internet Shopping Optimization Problem 2.7 Research Open Issues 3 Conclusions References A Comparison Between MFCC and MSE Features for Text-Independent Speaker Recognition Using Machine Learning Algorithms 1 Introduction 1.1 Database Description 1.2 Literature Reviewed 2 Proposed Methodology and Feature Extraction 2.1 Preprocessing 2.2 Mel Frequency Cepstral Coefficients 2.3 Multiband Spectral Entropy 2.4 Machine Learning Algorithms 2.5 Decision Trees 2.6 Random Forest 2.7 K-nearest Neighbors 2.8 Artificial Neural Networks 2.9 Preprocessing for Classification 2.10 Setup for Machine Learning Algorithms 2.11 Defined Experiments 3 Results 3.1 Speaker Recognition 3.2 Speaker Recognition in Men 3.3 Speaker Recognition in Women 3.4 Genre Recognition 4 Conclusions References Forecasting Based on Fuzzy Logic of the Level of Epidemiological Risk for the Mexican State of Tamaulipas 1 Introduction 2 Fuzzy Time Series Definitions 3 Proposed Forecasting Model 3.1 Universe of Discourse 3.2 Partition of the Universe of Discourse 3.3 Fuzzification 3.4 Fuzzy Logical Relationships 3.5 Defuzzification 3.6 Computational Implementation and Model Accuracy 4 Conclusions References Bio-inspired Flower Pollination Algorithm for the Optimization of a Monolithic Neural Network 1 Introduction 2 Artificial Neural Networks (ANNs) 2.1 Multilayer Perceptron (MLP) 3 Flower Pollination Algorithm (FPA) 4 Monolithic Neural Network Optimization with FPA 5 Simulation 5.1 ORL Database 5.2 Method: Manual Adjustment of Parameters (MAN-NN) 5.3 Neural Network Optimized with FPA (FPA-NN) 5.4 Statistical Test Between Methods 6 Conclusions References Rendezvous and Docking Control of Satellites Using Chaos Synchronization Method with Intuitionistic Fuzzy Sliding Mode Control 1 Introduction 2 Dynamic Behavior of Satellites 2.1 Chaos Dynamics of the Satellites 2.2 Synchronization of Two Satellites 3 Design of Controllers 3.1 Design of SMC 3.2 Designing Fundamental Concepts of FSMC and IFSMC 4 Numerical Simulation for Satellite Synchronization 4.1 Numerical Simulation Results for FSMC 4.2 Numerical Simulation Results for IFSMC Method 4.3 Numerical Simulation Results According to kf 5 Conclusions and Future Works References Optimizing a Convolutional Neural Network with a Hierarchical Genetic Algorithm for Diabetic Retinopathy Detection 1 Introduction 2 Basic Concepts 2.1 Neural Networks 2.2 Hierarchical Genetic Algorithms 3 Proposed Method 3.1 Pre-processing Applied to APTOS 2019 3.2 Genes of the Chromosomes 4 Experimental Results 5 Conclusions References Interval Type-3 Fuzzy Systems: A Natural Evolution from Type-1 and Type-2 Fuzzy Systems 1 Introduction 2 Interval Type-3 Fuzzy Sets 3 Footprint of Uncertainty 4 Interval Type-3 Fuzzy Systems 5 Simulation Results 6 Conclusions References A Comparative Study Between Bird Swarm Algorithm and Artificial Gorilla Troops Optimizer 1 Introduction 2 Literature Review 2.1 Optimization 2.2 Bird Swarm Algorithm 2.3 Artificial Gorilla Troops Optimizer 2.4 Optimization Problems Using BSA 2.5 Optimization Problems Using GTO 3 Mathematical Benchmark Functions 4 Results 5 Conclusions and Future Work References Particle Swarm Optimization of Convolutional Neural Networks for Diabetic Retinopathy Classification 1 Introduction 2 Basic Concepts 2.1 Convolutional Neural Networks 2.2 Particle Swarm Optimization 3 General Description of the Proposed Method 3.1 Description of PSO 3.2 Database 4 Experimental Results 4.1 Pre-processing #1 4.2 Pre-processing #2 5 Conclusions References