دانلود کتاب Innovations in Machine and Deep Learning: Case Studies and Applications (Studies in Big Data, 134)

فهرست مطالب :

Preface
Contents
Analytics-Oriented Applications
Recursive Multi-step Time-Series Forecasting for Residual-Feedback Artificial Neural Networks: A Survey
1 Introduction
2 Residual-Feedback ANNs: A Systematic Review
2.1 Systematic Review Planning and Execution
2.2 Overview of the Systematic Review Findings
3 The Existing Recursive Multi-step Forecast Strategy Solution
4 Limitation
5 Conclusions and Future Works
References
Feature Selection: Traditional and Wrapping Techniques with Tabu Search
1 Introduction
2 Related Work
3 Methodology
3.1 Data Description
3.2 Entropy-Based Feature Selection
3.3 Feature Selection Using Principal Component Analysis
3.4 Correlation-Based Feature Selection
4 Tabu Search
4.1 Initial Solution
4.2 Neighborhood
4.3 Objective Function
4.4 Memory Structures
5 Results
6 Discussion
7 Conclusions and Future Work
References
Pattern Classification with Holographic Neural Networks: A New Tool for Feature Selection
1 Introduction
2 Holographic Neural Networks
2.1 Basic Theory
2.2 Learning and Prediction Methods
2.3 red Explainability and Optimization of Holographic Models
3 Feature Selection with Holographic Neural Neworks
3.1 Previous Works
3.2 Pythagorean Membership Grades
4 Pattern Classification
4.1 Iris Dataset
4.2 red NIPS Feature Selection Challenge
5 red Conclusions and Future Works
References
Reusability Analysis of K-Nearest Neighbors Variants for Classification Models
1 Introduction
2 The K-Nearest Neighbors Algorithm
3 The Parameter K
4 Closeness Metrics
5 Analysis of KNN Variants
5.1 Heuristics for Class Assignment
5.2 Reduction of Dataset Records
5.3 Estimation of Dataset Variables
5.4 Discussion
6 Conclusions
References
Speech Emotion Recognition Using Deep CNNs Trained on Log-Frequency Spectrograms
1 Introduction
2 Literature Survey
2.1 Motivation
2.2 Contributions
3 Proposed Methodology
3.1 Data Augmentation
3.2 Extraction of Log-Frequency Spectrograms
3.3 Motivation Behind Using Spectrograms
3.4 Log-Frequency Spectrogram Extraction
3.5 Understanding What a Spectrogram Conveys
4 The Deep Convolutional Neural Network
4.1 Architecture
4.2 Training
5 Observations
5.1 Dataset Used
5.2 Performance Metrics Used
5.3 Results Obtained
5.4 Comparison Study
6 Conclusion
References
Text Classifier of Sensationalist Headlines in Spanish Using BERT-Based Models
1 Introduction
2 Background
2.1 Sensationalism
2.2 BERT-Based Models
3 Related Work
4 Dataset and Methods
4.1 Data Gathering and Data Labeling
4.2 Data Analysis
4.3 Model Generation and Fine-Tuning
5 Results
6 Conclusion
References
Arabic Question-Answering System Based on Deep Learning Models
1 Introduction
2 Natural Language Processing (NLP)
2.1 Difficulties in NLP
2.2 Natural Language Processing Phases
3 Question Answer System
3.1 Usage Deep Learning Models in Questions Answering System
3.2 Different Questions Based on Bloom’s Taxonomy
3.3 Question-Answering System Based on Types
3.4 Wh-Type Questions (What, Which, When, Who)
4 List-Based Questions
5 Yes/No Questions
6 Causal Questions [Why or How]
7 Hypothetical Questions
8 Complex Questions
8.1 Question Answering System Issues
9 Arabic Language Overview
9.1 Arabic Language Challenges
10 Related Work
11 Proposed Methodology
11.1 Recurrent Neural Networks (RNNs)
11.2 Long Short-Term Memory (LSTM)
11.3 Gated Recurrent Unit (GRU)
12 Prepare the Dataset
12.1 Collecting Data
13 Data Preprocessing
14 Results and Discussion
15 Conclusion and Future Work
References
Healthcare-Oriented Applications
Machine and Deep Learning Algorithms for ADHD Detection: A Review
1 Introduction
2 Research Methodology
3 Related Work
3.1 Machine Learning Approaches
3.2 Deep Learning Approaches
4 Approaches for ADHD Detection Using AI Algorithms
4.1 Machine Learning-Based Approaches
4.2 Deep Learning-Based Approaches
5 Datasets for ADHD Detection
5.1 Hyperaktiv
5.2 Working Memory and Reward in Children with and Without ADHD
5.3 Working Memory and Reward in Adults
5.4 Eeg Data for ADHD
6 Machine Learning and Deep Learning Classifiers for ADHD Detection
7 Trends and Challenges
7.1 New Types of Sensors or Biosensors
7.2 Multi-Modal Detection and/or Diagnosis of ADHD
7.3 The Use of Biomarkers as Variables for Diagnosis
7.4 Interpretability
7.5 Building of Standardized and Accurate Public Datasets
7.6 Different Classification Techniques
8 Conclusion
References
Mosquito on Human Skin Classification Using Deep Learning
1 Introduction
2 Literature Review
3 Methodology
3.1 Dataset Description
3.2 Deep Convolutional Neural Networks and Transfer Learning
3.3 Hyperparameter Tuning
3.4 Proposed Workflow
4 Experiments and Results
5 Conclusion and Future Work
References
Analysis and Interpretation of Deep Convolutional Features Using Self-organizing Maps
1 Introduction
2 Materials
2.1 Convolutional Neural Networks
2.2 Self-organizing Maps
3 Proposed Method
3.1 Stage A: Training of CNN
3.2 Stage B: Extraction of Features
3.3 Stage C: SOM Training
3.4 Stage D: Analysis and Interpretation
4 Application Example
4.1 Experimental Setup
4.2 Result Analysis
5 Conclusions
References
A Hybrid Deep Learning-Based Approach for Human Activity Recognition Using Wearable Sensors
1 Introduction
2 Literature Analysis
3 OPPORTUNITY Dataset
4 MHEALTH Dataset
5 HARTH Dataset
6 Materials and Methods
6.1 Some Preliminaries
6.2 Basic Architecture of CNN
7 Long-Short Term Memory (LSTM)
7.1 Working Principle of LSTM
8 Proposed Model Architecture
9 Dataset Description
9.1 MHEALTH Dataset
9.2 OPPORTUNITY Dataset
9.3 HARTH Dataset
10 Experimental Results
10.1 Evaluation Metrics Used
10.2 Results Analysis on MHEALTH Dataset
10.3 Results Analysis on OPPORTUNITY Dataset
10.4 Results Analysis on HARTH Dataset
10.5 Result Summary and Comparison
11 Conclusion and Future Works
References
Predirol: Predicting Cholesterol Saturation Levels Using Big Data, Logistic Regression, and Dissipative Particle Dynamics Simulation
1 Introduction
2 Related Works
2.1 Models for the Simulation of Fluids
2.2 Data Mining Application for Prevention of Cardiovascular Diseases
2.3 Comparative Analysis
3 PREDIROL Architecture
3.1 Big Data Model
3.2 Cholesterol Saturation Level Prediction Module
3.3 Cholesterol Levels Simulation Module with Dissipative Particle Dynamics
4 Case Study: Prediction of Cholesterol Levels of a Hospital Patients
5 Conclusions and Future Work
References
Convolutional Neural Network-Based Cancer Detection Using Histopathologic Images
1 Introduction
2 Image Processing Techniques
2.1 Statistical-Based Algorithms
2.2 Learning-Based Algorithms
2.3 Hyper-Parameters of CNN
2.4 Evaluation Metrics
2.5 Implementation
3 Stage 3: CNN Algorithm Training
3.1 Model Training Phase
3.2 Model Optimization Phase
4 Conclusion
References
Artificial Neural Network-Based Model to Characterize the Reverberation Time of a Neonatal Incubator
1 Introduction
2 Materials and Methods
2.1 Artificial Neural Networks Using the Levenberg–Marquardt Algorithm
3 Results
3.1 Data Analysis
3.2 Artificial Neural Network-Based Model Training
4 Conclusions
References
A Comparative Study of Machine Learning Methods to Predict COVID-19
1 Introduction
2 Related Works
3 Background
3.1 Covid-19
3.2 Machine Learning
4 Materials and Methods
4.1 Dataset Pre-processing
4.2 Machine Learning Models
5 Results and Discussions
6 Conclusions
References
Sustainability-Oriented Applications
Multi-product Inventory Supply and Distribution Model with Non-linear CO2 Emission Model to Improve Economic and Environmental Aspects of Freight Transportation
1 Introduction
2 Literature Review and Contributions
3 Development of the Integrated Routing Model
3.1 Inventory Planning with Non-deterministic Demand and Multiple Products
3.2 Non-linear Emission for Heterogeneous Fleet
3.3 Association of Variables
4 Assessment of the Model
4.1 Numerical Data and Solving Method
4.2 Analysis of Results
5 Future Work
6 Statement
References
Convolutional Neural Networks for Planting System Detection of Olive Groves
1 Background
1.1 Evolution of Production Techniques in Olive Groves
1.2 Current Situation of Modern Olive Cultivation Systems
1.3 Application of Remote Sensing Techniques for Image Analysis
1.4 Scope of the Present Chapter
2 Materials and Experimental Methods
2.1 Area of Study and Image Acquisition
2.2 Methodology
3 Results and Discussion
4 Conclusions and Future Lines
References
A Conceptual Model for Analysis of Plant Diseases Through EfficientNet: Towards Precision Farming
1 Introduction
2 Related Study
3 Deep Learning in Keras
4 Overview of Plant Diseases
5 Materials and Methods
5.1 Dataset Used in the Study
5.2 Overview of Convolutional Neural Network Models
5.3 Overview of EfficientNet
5.4 B0 to B7 Variants of EffcientNet
6 Proposed Methodology for Plant Disease Detection from Leaf Images
6.1 Experimental Setup
6.2 Training
7 Results and Discussion
7.1 Evaluation of Model
7.2 Image Analysis
8 Conclusion
References
Ginger Disease Detection Using a Computer Vision Pre-trained Model
1 Introduction
2 Related Work
3 Data Preparation
4 Pre-trained Model Description
5 Methodology
6 Hyper-Parameter Setting
7 Experimental Result
8 Conclusion
References
Anomaly Detection in Low-Cost Sensors in Agricultural Applications Based on Time Series with Seasonal Variation
1 Introduction
2 Related Work
3 Problem Statement
4 Anomaly Detection Methodology
4.1 Methodology Basis
4.2 Anomaly Detector Enhancements
5 Evaluation of the Proposed Approach
5.1 Data Generation
5.2 Experimental Results
6 Conclusion
References
Coconut Tree Detection Using Deep Learning Models
1 Introduction
2 Related Studies
3 Proposed Work
3.1 Datasets
4 Training and Classification
4.1 Model Selection
4.2 Evaluation Metrics
5 Experiments and Results
5.1 Model Trained with Low-Resolution Images
5.2 Model Trained with High-Resolution Images
5.3 Graphical User Interface—GUI
6 Conclusion and Future Work
References
Hybrid Neural Network Meta-heuristic for Solving Large Traveling Salesman Problem
1 Introduction
2 Clustering for Dimensionality Reduction
2.1 Self-organizing Maps
2.2 Reduction of Location Data
3 Structure of the Hybrid ANN-CW-LS Meta-heuristic
4 Results and Assessment of Performance
4.1 Test Data and Experiment Settings
4.2 Test of Average Performance
4.3 Test of Best Performance
4.4 Speed Performance Versus Dimensionality Reduction
5 Conclusions and Future Work
References