Metaheuristics in Machine Learning: Theory and Applications (Studies in Computational Intelligence, 967)

✍ Scribed by Diego Oliva (editor), Essam H. Houssein (editor), Salvador Hinojosa (editor)

• Publisher: Springer
• Year: 2021
• Tongue: English
• Leaves: 765
• Edition: 1st ed. 2021
• Category: Library

✦ Synopsis

This book is a collection of the most recent approaches that combine metaheuristics and machine learning. Some of the methods considered in this book are evolutionary, swarm, machine learning, and deep learning. The chapters were classified based on the content; then, the sections are thematic. Different applications and implementations are included; in this sense, the book provides theory and practical content with novel machine learning and metaheuristic algorithms.

The chapters were compiled using a scientific perspective. Accordingly, the book is primarily intended for undergraduate and postgraduate students of Science, Engineering, and Computational Mathematics and is useful in courses on Artificial Intelligence, Advanced Machine Learning, among others. Likewise, the book is useful for research from the evolutionary computation, artificial intelligence, and image processing communities. 

✦ Table of Contents

Preface
Introduction
Contents
Cross Entropy Based Thresholding Segmentation of Magnetic Resonance Prostatic Images Using Metaheuristic Algorithms
1 Introduction
1.1 Applied Metaheuristic Algorithms
2 Moth-Flame Optimizer Algorithm
3 Sine Cosine Optimization Algorithm
4 Sunflower Optimization Algorithm
5 Image Segmentation Using Minimum Cross Entropy
6 Results of the Experiments
6.1 Experimental Setup
6.2 Metrics and Experimental Results
7 Conclusions
Appendix: MRIs Detailed Series Header Information
References
Hyperparameter Optimization in a Convolutional Neural Network Using Metaheuristic Algorithms
1 Introduction
2 Convolutional Neural Networks
2.1 Artificial Neuron
2.2 Artificial Neural Network
2.3 Training
2.4 Convolutional Neural Network Architecture
3 Hyperparameters
4 Metaheuristic Algorithms
4.1 Ant Lion Optimization (ALO)
4.2 Artificial Bee Colony (ABC)
4.3 Bat Algorithm (BA)
4.4 Particle Swarm Optimization
5 The General Procedure for Hyperparameter Optimization in a Convolutional Neural Network Using Metaheuristic Algorithms
6 Experimental Result
7 Conclusion
References
Diagnosis of Collateral Effects in Climate Change Through the Identification of Leaf Damage Using a Novel Heuristics and Machine Learning Framework
1 Introduction
2 Materials and Methods
2.1 Data Acquisition
2.2 Frequency Histogram
2.3 Brightness Reduction
2.4 Contrast Enhancement
2.5 Segmentation
2.6 Otsu Method Thresholding
2.7 Database Creation
2.8 Classification
3 Experiments and Results
4 Analysis and Discussion
5 Conclusions and Future Research
References
Feature Engineering for Machine Learning and Deep Learning Assisted Wireless Communication
1 Introduction
2 Feature Engineering for Machine Learning and Deep Learning Assisted Wireless Communication System
3 Feature Transformation
3.1 Feature Construction
3.2 Feature Extraction
3.3 Feature Transformation Application for Automatic Modulation Classification
4 Features Selection
4.1 Feature Analysis and Evaluation
4.2 Feature Engineering Application for Path-Loss Prediction in Wireless Communication
4.3 Feature Engineering-Based Path-Loss Prediction
5 Solutions and Recommendations
6 Conclusion
References
Genetic Operators and Their Impact on the Training of Deep Neural Networks
1 Introduction
2 Preliminary Concepts
2.1 Genetic Algorithm
2.2 Artificial Neural Networks
2.3 Direct and Indirect Encoding
3 Problem Definition
3.1 The Maze Problem
3.2 Neural Network Architecture
3.3 Genetic Algorithm
4 Experimental Results
4.1 GA Variation 1 Analysis
4.2 GA Variation 2 Analysis
4.3 GA Variation 3 Analysis
4.4 GA Variation 4 Analysis
4.5 Statistical Validation
5 Conclusions
References
Implementation of Metaheuristics with Extreme Learning Machines
1 Introduction
2 Extreme Learning Machines (ELM)
3 Swarm Intelligence Metaheuristics
3.1 Particle Swarm Optimization (PSO)
3.2 Grey Wolf Optimization
3.3 Artificial Bee Colony (ABC)
4 Metaheuristics in the Extreme Learning Machine Method
5 Results
6 Conclusions
Appendix
References
Architecture Optimization of Convolutional Neural Networks by Micro Genetic Algorithms
1 Deep Learning and Neuroevolution
2 Deep CNN, GA and Neuroevolution Approaches
2.1 Deep Convolutional Neural Networks
2.2 Genetic Algorithms
2.3 Neuroevolution
3 The Micro Genetic Algorithm CNN Framework Proposal
4 Experiments and Results
5 Conclusion
References
Optimising Connection Weights in Neural Networks Using a Memetic Algorithm Incorporating Chaos Theory
1 Introduction
2 Feed-Forward Neural Networks
3 Swarm Intelligence
4 Memetic Algorithms
5 Colonial Competitive Algorithm
6 Proposed Algorithm
6.1 Representation
6.2 Cost Function
6.3 Back-Propagation Algorithm as a Local Search Operator
6.4 Chaos-Enhanced Memetic Algorithm
7 Experimental Results
7.1 Sigmoid Function
7.2 Cosine Function
7.3 Sine1 Function
7.4 Sine2 Function
8 Conclusions
References
A Review of Metaheuristic Optimization Algorithms in Wireless Sensor Networks
1 Introduction
2 Metaheuristic Intelligence Optimization and Evolutionary Algorithms
2.1 Evolutionary Algorithms
2.2 Swarm Intelligence-Based Algorithms
2.3 Bio-Inspired Algorithms
2.4 Physics and Chemistry-Based Algorithms
2.5 Other Algorithms
3 Wireless Sensor Networks
4 Application of Metaheuristic Algorithms in Wireless Sensor Networks
4.1 Optimization Algorithms
4.2 Deployment in Wireless Sensor Networks
4.3 Localization in Wireless Sensor Networks
4.4 Sink Node Placement and Energy Consumption in Wireless Sensor Networks
5 Conclusion
References
A Metaheuristic Algorithm for Classification of White Blood Cells in Healthcare Informatics
1 Introduction
2 Cognitive Computing Concept
3 Neural Networks Concepts
3.1 Convolutional Neural Network
4 Metaheuristic Algorithm Proposal
5 Results and Discussion
6 Future Research Directions
7 Conclusions
References
Multi-level Thresholding Image Segmentation Based on Nature-Inspired Optimization Algorithms: A Comprehensive Review
1 Introduction
2 Nature-Inspired Optimization Algorithms
2.1 Evolutionary-Based Techniques
2.2 Bio-Inspired Based Algorithms
2.3 Physics and Chemistry Based Algorithms
2.4 Other Algorithms
3 Segmentation Techniques
3.1 Threshold-Based Methods
3.2 Region-Based Methods
3.3 Edge Detection Methods
3.4 Clustering Methods
3.5 Neural Network Based Methods
3.6 Partial Differential Equation Methods
4 Segmentation Quality Assessment Parameters
4.1 Peak Signal to Noise Ratio
4.2 Structural Similarity Index Measure
4.3 Feature Similarity Index Measure
4.4 Mean Square Error
4.5 Quality Index Based on Local Variance
5 Nature-Inspired Optimization Algorithms in Image Multi-thresholding
6 Open Problems and Challenges
7 Conclusion and Future Research Issues
References
Hybrid Harris Hawks Optimization with Differential Evolution for Data Clustering
1 Introduction
2 Related Works
3 Preliminaries
3.1 Cluster Analysis
3.2 External Measure
4 H-HHO: Hybrid Harris Hawks Optimization with Differential Evolution
4.1 Harris Hawks Optimization Algorithm
5 Experiments Results and Discussions
5.1 Experimental Setup
5.2 Comparisons for Data Clustering
6 Conclusion and Future Work
References
Variable Mesh Optimization for Continuous Optimization and Multimodal Problems
1 Introduction
2 Variable Mesh Optimization for Continuous Optimization Problems
3 VMO with Niching Strategies for Multimodal Problems
3.1 Niching VMO
3.2 Generic Niching Framework for VMO
4 Empirical Assessment
5 Discussion and Conclusions
References
Traffic Control Using Image Processing and Deep Learning Techniques
1 Introduction
2 Background
2.1 Methodologies for Object Identification and Image Processing
2.2 Methodologies for Urban Mobility Systems
3 Proposal
3.1 Prerequisites
3.2 Green Light On—Minimum Time Analysis
3.3 Equations
4 Experiments
4.1 Simulation Using Random Data
4.2 Real Situation
5 Results
5.1 Results Using Random Data
5.2 Real Situation
6 Conclusion
References
Drug Design and Discovery: Theory, Applications, Open Issues and Challenges
1 Introduction
2 Background
3 Different Techniques Overview Used in Drug Design and Discovery
4 Drug Design and Discovery Overview
5 Applications of Drug Design and Discovery
6 Open Issues and Challenges
7 Conclusions
References
Thresholding Algorithm Applied to Chest X-Ray Images with Pneumonia
1 Introduction
2 The Whale Optimization Algorithm
2.1 Exploitation Phase
2.2 Exploration Phase
3 Image Multilevel Thresholding
3.1 Otsu's Method Based in Between-Class Variance
3.2 Kapur's Method Based in Entropy
4 Automatic Detection of Thresholds Values Using WOA with Kapur and Otsu as Objective Functions
4.1 Dataset Description
4.2 Experiments Details
4.3 Metrics
5 Experimental Results
5.1 Results of Otsu's Objective Function
5.2 Results of Kapur's Objective Function
6 Conclusions
References
Artificial Neural Networks for Stock Market Prediction: A Comprehensive Review
1 Introduction
2 Artificial Neural Networks and Stock Price Prediction
2.1 Artificial Neural Networks
2.2 Training the ANNs
3 Stock Market Prediction: Description and Need
3.1 Stock Market Prediction Techniques: A Survey
4 Discussion
4.1 Publication Years
4.2 Prediction Techniques
4.3 Data Sets
4.4 Performance Evaluation
4.5 Prediction Target
5 Conclusion
References
Image Classification with Convolutional Neural Networks
1 Introduction
2 Image Classification
3 Classification Metrics
3.1 Binary Classifier Metrics
3.2 Multi-class Classifier Metrics
4 Neural Nets
4.1 Nodes
4.2 Tensors and Data Batches
4.3 Activation Function
4.4 Layers
4.5 Supervised Learning
4.6 Learning Rate
5 Convolutional Neural Networks
5.1 Convolutional Layer
5.2 Pooling Layer
5.3 Fully Connected Layers
6 Training CNN
7 Results
8 Conclusions
References
Applied Machine Learning Techniques to Find Patterns and Trends in the Use of Bicycle Sharing Systems Influenced by Traffic Accidents and Violent Events in Guadalajara, Mexico
1 Introduction
2 Methodology
3 Business Comprehension
4 Data Comprehension
5 Data Preparation
6 Data Modelling and Evaluation
7 Conclusions and Future Works
References
Machine Reading Comprehension (LSTM) Review (State of Art)
1 Introduction
2 Machine Learning
2.1 Machine Comprehension (MC)
2.2 Artificial Intelligence
2.3 Machine Learning
2.4 Neural Networks
2.5 Deep Learning
3 Classifications
3.1 Natural Language Processing (NLP)
3.2 Machine Reading Comprehension (MRC)
3.3 Recurrent Neural Network RNN
3.4 Long Short-Term Memory Networks (LSTM)
3.5 Recent Research
3.6 Success Cases
4 Conclusion
References
A Survey of Metaheuristic Algorithms for Solving Optimization Problems
1 Introduction
2 Optimization Problems Types
2.1 Single Objective Optimization
2.2 Multi-objective Optimization
2.3 Combinatorial Optimization Problems
2.4 Continuous Optimization Problem
3 MOA for the Single Objective
3.1 Engineering Design Problems
3.2 Feature Selection
3.3 Others
4 MOA for Multiobjective Optimization Problems
4.1 Feature Selection Optimization Problem
4.2 Engineering Optimization Problem
4.3 Other Optimization Problems
5 Open Issues and Challenges and Future Trends
5.1 Open Issues (Problems)
5.2 Future Trends
6 Conclusion
References
Integrating Metaheuristic Algorithms and Minimum Cross Entropy for Image Segmentation in Mist Conditions
1 Introduction
2 Metaheuristic Algorithms for Optimization
2.1 PSO
2.2 ABC
2.3 SCA
2.4 GWO
2.5 HHO
3 Segmentation and Thresholding Methods
3.1 Minimum Cross Entropy
4 Comparison Criteria
4.1 MSE
4.2 PSNR
4.3 SSIM
4.4 FSIM
4.5 Average and Standard Deviation as Statistical Measures
5 Methodology
5.1 Software Processing
6 Experimental Results
7 Conclusions
Appendix
References
Machine Learning Application for Particle Physics: Mexico's Involvement in the Hyper-Kamiokande Observatory
1 Introduction
2 Mexico and the Hyper-Kamiokande Experiment
2.1 The Participation of Mexico in Hyper-K
2.2 The Hyper-Kamiokande Experiment
2.3 Supercomputing Facilities in Mexico to Be Used on Hyper-K
3 Machine Learning Theory
3.1 Neural Networks
3.2 Training NN with Gradient-Based Method
4 CNN Application for Particle Identification in the Hyper-Kamiokande Experiment
4.1 From Monte Carlo Simulation to Image-Like Data
4.2 CNN Implementation
4.3 Results and Discussion
5 Conclusions and Future Work
References
A Novel Metaheuristic Approach for Image Contrast Enhancement Based on Gray-Scale Mapping
1 Introduction
2 Image Contrast Enhancement Based on Gray-Scale Mapping
3 Moth Swarm Algorithm
3.1 Initialization
3.2 Recognition Phase
3.3 Transversal Flight
3.4 Celestial Navigation
4 MSA-Based Image Contrast Enhancement via Gray-Levels Mapping
4.1 Objective Function
4.2 Penalty Function
5 Experimental Results
5.1 Standard Test Images
5.2 Low Contrast Test Images
6 Conclusions
References
Geospatial Data Mining Techniques Survey
1 Introduction
2 Review of Concepts
3 Definition of Geospatial Data Mining
4 Techniques of Geospatial Data Mining
4.1 Spatial Class Identification
4.2 Spatial Outlier Detection
4.3 Spatial Association Rules
4.4 Spatial Cluster Analysis
4.5 Trends and Deviations Analysis
5 Geospatial Data Mining Approaches
6 Conclusions
References
Integration of Internet of Things and Cloud Computing for Cardiac Health Recognition
1 Introduction
2 Exploration of ECG Analysis
2.1 ECG Data Acquisition
2.2 Filtering
2.3 Heartbeat Detection
2.4 Heartbeat Segmentation
2.5 Feature Extraction
2.6 Feature Selection
2.7 Classification
3 IoT Framework for ECG Monitoring System
3.1 IoT Topology
3.2 IoT Structure
3.3 IoT Platform
4 Major Issues for ECG Monitoring System
4.1 Monitoring Device
4.2 Energy Efficiency
4.3 Signal Quality
4.4 Data
4.5 System Integration
5 Conclusion
References
Combinatorial Optimization for Artificial Intelligence Enabled Mobile Network Automation
1 Introduction
2 Network Automation Use-Cases
2.1 Physical Cell ID (PCI) Assignment
2.2 PRACH Configuration
2.3 Mobility Robustness Optimization
2.4 Mobility Load Balancing
3 Combinatorial Optimization Algorithms
3.1 Preliminaries
3.2 Heuristics and Metaheuristics
3.3 Trajectory Metaheuristics
3.4 Population Based Metaheuristics
4 Applications in Mobile Network Automation
4.1 Local Search
4.2 Simulated Annealing
4.3 Tabu Search
4.4 Evolutionary Computation
5 Greedy Heuristics for PCI Assignment: A Case Study
5.1 Model
5.2 Algorithms
5.3 Results
6 Conclusions and Future Work
6.1 Conclusions
6.2 Future Work
References
Performance Optimization of PID Controller Based on Parameters Estimation Using Meta-Heuristic Techniques: A Comparative Study
1 Introduction
2 PID Controller
3 Tuning Parameters of PID Controllers Using Meta-Heuristic Algorithms
4 PSO, SCA, ASCA-PSOand EVOA Algorithms
4.1 Particle Swarm Optimization (PSO) Algorithm [11]
4.2 Sine–Cosine Optimization (SCA) Algorithm [9]
4.3 Hybrid SCA and PSO Algorithm (ASCA-PSO) [18]
4.4 Egyptian Vulture Optimization Algorithm (EVOA) [22]
5 Experimental Results
5.1 Liquid Level System
5.2 3rd and 4th Order Systems Process
6 Conclusion and Future Work
References
Solar Irradiation Changes Detection for Photovoltaic Systems Through ANN Trained with a Metaheuristic Algorithm
1 Introduction
2 Maximum Power Point for Photovolatic Systems
3 Artificial Neural Networks
4 Earthquake Algorithm
4.1 Earthquake Algorithm as ANN Tuning Method
4.2 MATLAB Implementation of the EA for ANN Training
5 Case Study
5.1 ANN for Solar Irradiation Changes Detection
5.2 Simulink Implementation for Validation
6 Results
6.1 Quantification
7 Conclusion
References
Genetic Algorithm Based Global and Local Feature Selection Approach for Handwritten Numeral Recognition
1 Introduction
2 Literature Survey
3 Collection and Pre-processing of Numeral Databases
3.1 Pre-processing
3.2 Extraction of Individual Digits
3.3 Image Resizing
3.4 Image Binarization
3.5 Image Thinning
4 Design of Feature Set
4.1 Global Features
4.2 Local Distance Based Features
4.3 Selection of Optimal Feature Subset Using GA
5 Experimental Analysis
6 Conclusion
References