Machine Learning, Big Data, and IoT for Medical Informatics (Intelligent Data-Centric Systems: Sensor Collected Intelligence)

Front matter
Copyright
Contributors
Preface
	Outline of the book and chapter synopses
	Special acknowledgments
Predictive analytics and machine learning for medical informatics: A survey of tasks and techniques
	Introduction: Predictive analytics for medical informatics
		Overview: Goals of machine learning
		Current state of practice
		Key task definitions
			Diagnosis
			Predictive analytics
			Therapy recommendation
			Automation of treatment
			Other tasks in integrative medicine
		Open research problems
			Learning for classification and regression
			Learning to act: Control and planning
			Toward greater autonomy: Active learning and self-supervision
	Background
		Diagnosis
			Diagnostic classification and regression tasks
			Diagnostic policy-learning tasks
			Active, transfer, and self-supervised learning
		Predictive analytics
			Prediction by classification and regression
			Learning to predict from reinforcements and by supervision
			Transfer learning in prediction
		Therapy recommendation
			Supervised therapy recommender systems
		Automation of treatment
			Classification and regression-based tasks
			RL for automation
			Active learning in automation
		Integrating medical informatics and health informatics
			Classification and regression tasks in HMI
			Reinforcement learning for HMI
			Self-supervised, transfer, and active learning in HMI
	Techniques for machine learning
		Supervised, unsupervised, and semisupervised learning
			Shallow
			Deep
		Reinforcement learning
			Traditional
			Deep RL
		Self-supervised, transfer, and active learning
			Traditional
			Deep
	Applications
		Test beds for diagnosis and prognosis
			New test beds
		Test beds for therapy recommendation and automation
			Prescriptions
			Surgery
	Experimental results
		Test bed
		Results and discussion
	Conclusion: Machine learning for computational medicine
		Frontiers: Preclinical, translational, and clinical
		Toward the future: Learning and medical automation
	References
Geolocation-aware IoT and cloud-fog-based solutions for healthcare
	Introduction
	Related work
		Health monitoring system with cloud computing
		Health monitoring system with fog computing
		Health monitoring system with cloud-fog computing
	Proposed framework
		Health data analysis
		Geospatial analysis for medical facility
			Overlay analysis to obtain nearest medical facilities
			Shortest path to reach nearest medical centers
		Delay and power consumption calculation
	Performance evaluation
	Conclusion and future work
	References
Machine learning vulnerability in medical imaging
	Introduction
	Computer vision
	Adversarial computer vision
	Methods to produce adversarial examples
	Adversarial attacks
	Adversarial defensive methods
	Adversarial computer vision in medical imaging
	Adversarial examples: How to generate?
	Conclusion
	Acknowledgment
	References
Skull stripping and tumor detection using 3D U-Net
	Introduction
		Previous work
	Overview of U-net architecture
		3D U-net
			Batch normalization
			Activation function
			Pooling
			Padding
			Optimizer
	Materials and methods
		Dataset
		Implementation
	Results
		Experimental result
			Dice coefficient
			Accuracy
			Intersection over Union (IoU)
		Quantitative result
		Qualitative result
	Conclusion
	References
Cross color dominant deep autoencoder for quality enhancement of laparoscopic video: A hybrid deep learning an
	Introduction
	Range-domain filtering
	Cross color dominant deep autoencoder (C2D2A) leveraging color spareness and saliency
		Evolution of DCM through C2D2A
		Inclusion of DCM into principal flow of bilateral filtering
	Experimental results
	Conclusion
	Acknowledgments
	References
Estimating the respiratory rate from ECG and PPG using machine learning techniques
	Introduction
		Motivation
		Background
	Related work
	Methods
		Data
		Steps
		RR signal extraction
		Machine learning
	Experimental results
	Discussion and conclusion
	Acknowledgments
	References
Machine learning-enabled Internet of Things for medical informatics
	Introduction
		Healthcare Internet of Things
			H-IoT architecture
			Three-tier H-IoT architecture
	Applications and challenges of H-IoT
		Applications of H-IoT
			Fitness tracking
			Neurological disorders
			Cardio vascular disorders
			Ambient-assisted living
		Challenges of H-IoT system
			QoS improvement
			Scalability challenges
	Machine learning
		Machine learning advancements at the application level of H-IoT
		Machine learning advancements at network level of H-IoT
	Future research directions
		Novel applications of ML in H-IoT
			Real-time monitoring and treatment
			Training for professionals
			Advanced prosthetics
		Research opportunities in network management
			Channel access
			Dynamic data management
			Fully autonomous operation
			Security
	Conclusion
	References
Edge detection-based segmentation for detecting skin lesions
	Introduction
	Previous works
	Materials and methods
		Elitist-Jaya algorithm
		Otsus method
	Proposed method
		Image preprocessing
		Edge detection
	Experiment and results
		Dataset
		Evaluation metrics
		Results and discussion
		Statistical analysis
	Conclusion
	References
A review of deep learning approaches in glove-based gesture classification
	Introduction
	Data gloves
		Early and commercial data gloves
		Sensing mechanism in data gloves
			Fiber-optic sensors
			Conductive strain sensors
			Inertial sensors
	Gesture taxonomies
	Gesture classification
		Classical machine learning algorithms
			K-nearest neighbor
			Support vector machine (SVM)
			Decision tree
			Artificial neural network (ANN)
			Probabilistic neural network (PNN)
		Glove-based gesture classification with classical machine learning algorithms
		Deep learning
			Convolutional neural network (CNN)
			Recurrent neural network (RNN)
		Glove-based gesture classification using deep learning
	Discussion and future trends
	Conclusion
	References
An ensemble approach for evaluating the cognitive performance of human population at high altitude
	Introduction
	Methodology
		Data collection
		Data processing and feature selection
		Differential expression analyses
		Association rule mining
		Experimental set-up
	Results and discussion
		Differential analyses-Cognitive and clinical features
		Discovered associative rules
		Discussion
	Future opportunities
	Conclusions
	Acknowledgment
	References
Machine learning in expert systems for disease diagnostics in human healthcare
	Introduction
	Types of expert systems
	Components of an expert system
	Techniques used in expert systems of medical diagnosis
	Existing expert systems
	Case studies
		Cancer diagnosis using rule-based expert system
		Alzheimers diagnosis using fuzzy-based expert systems
			Algorithm of fuzzy inference system
	Significance and novelty of expert systems
	Limitations of expert systems
	Conclusion
	Acknowledgment
	References
An entropy-based hybrid feature selection approach for medical datasets
	Introduction
		Deficiencies of the existing models
		Chapter organization
	Background of the present research
		Feature selection (FS)
	Methodology
		The entropy based feature selection approach
			Equi-class distribution of instances
			Splitting the dataset D into subsets: D1, D2, and D3
	Experiment and experimental results
		Experiment using suggested feature selection approach
	Discussion
		Performance analysis of the suggested feature selection approach
	Conclusions and future works
	Conflict of interest
	Appendix A
		Explanation on entropy-based feature extraction approach
	References
Machine learning for optimizing healthcare resources
	Introduction
	The state of the art
		Resource management
		Impact on peoples health
		Exit strategies
	Machine learning for health data analysis
	Feature selection techniques
		Filter approach
			Correlation based
			Information based
			Consistency based
			Distance based
		Wrapper approach
			Classic search algorithms
				Greedy search
				Best first search
			Metaheuristic algorithms
				Genetic algorithms
				Particle swarm optimization
				Ant colony optimization
				Artificial bee colony
				Grey wolf optimization
				Artificial immune system algorithms
				Gravitational search optimization
		Embedded approach
	Machine learning classifiers
		One-class vs. multiclass classification
		Supervised vs. unsupervised learning
	Case studies
		Experimental setup
		Case study 1: Diabetes data analysis
	Case study 2: COVID-19 data analysis
	Summary and future directions
	References
Interpretable semisupervised classifier for predicting cancer stages
	Introduction
	Self-labeling gray box
	Data preparation
	Experiments and discussion
		Influence of clinical and proteomic data on the prediction of cancer stage
		Influence of unlabeled data on the prediction of cancer stage
		Influence of unlabeled data on the prediction of cancer stage for rare cancer types
	Conclusions
	Acknowledgments
	References
Applications of blockchain technology in smart healthcare: An overview
	Introduction
		Comparison to other surveys
	Blockchain overview
		Key requirements
	Proposed healthcare monitoring framework
	Blockchain-enabled healthcare applications
	Potential challenges
	Concluding remarks
	Declaration of competing interest
	References
Prediction of leukemia by classification and clustering techniques
	Introduction
	Motivation
	Literature review
	Description of proposed system
		Introduction and related concepts
		Framework for the proposed system
			Support vector machine
			K-nearest neighbor
			K-means clustering
			Fuzzy c-means clustering
	Simulation results and discussion
	Conclusion and future directions
	References
Performance evaluation of fractal features toward seizure detection from electroencephalogram signals
	Introduction
	Fractal dimension
		Katz fractal dimension
		Higuchi fractal dimension
		Petrosian fractal dimension
	Dataset
	Experiments
	Results and discussion
	Conclusion
	Acknowledgments
	References
Integer period discrete Fourier transform-based algorithm for the identification of tandem repeats in the DNA ...
	Introduction
	Related work
	Algorithm for detection of TRs
		DNA sequences
		Numerical mapping
		Short time integer period discrete Fourier transform
		Thresholding
		Verification of the detected candidate TRs
	Performance analysis of the proposed algorithm
	Conclusion
	References
A blockchain solution for the privacy of patients medical data
	Introduction
	Stakeholders of healthcare industry
		Patients
		Pharmaceutical companies
		Healthcare providers (doctors, nurses, hospitals, nursing homes, clinics, etc.)
		Government
		Insurance companies
	Data protection laws for healthcare industry
	Medical data management
	Issues and challenges of healthcare industry
	Blockchain technology
		Features of blockchain
		Types of blockchain
		Working of blockchain
	Blockchain applications in healthcare
	Blockchain-based framework for privacy protection of patients data
	Conclusion
	References
A novel approach for securing e-health application in a cloud environment
	Introduction
		Contribution
	Motivation
		Related works
		Challenges
	Proposed system
	Conclusion
	References
An ensemble classifier approach for thyroid disease diagnosis using the AdaBoostM algorithm
	Introduction
	Data analytics
	Machine learning
	Approaching ensemble learning
	Understanding bagging
	Exploring boosting
	Discovering stacking
		Machine learning applications for healthcare analytics
		Machine learning-based model for disease diagnosis
		Machine learning-based algorithms to identify breast cancer
		Convolutional neural networks to detect cancer cells in brain images
		Machine learning techniques to detect prostate cancer in Magnetic resonance imaging
		Classification of respiratory diseases using machine learning
		Parkinsons disease diagnosis with machine learning-based models
	Processing drug discovery with machine learning
		Analyzing clinical data using machine learning algorithms
		Predicting thyroid disease using ensemble learning
		Machine learning-based applications for thyroid disease classification
		Preprocessing the dataset
		AdaBoostM algorithm
	Conclusion
	References
A review of deep learning models for medical diagnosis
	Motivation
	Introduction
	MRI Segmentation
	Deep learning architectures used in diagnostic brain tumor analysis
		Convolutional neural networks or convnets
		Stacked autoencoders
		Deep belief networks
		2D U-Net
		3D U-Net
		Cascaded anisotropic network
	Deep learning tools applied to MRI images
	Proposed framework
	Conclusion and outlook
	Future directions
	References
Machine learning in precision medicine
	Precision medicine
	Machine learning
	Machine learning in precision medicine
		Detection and diagnosis of a disease
		Prognosis of a disease
		Discovery of biomarkers and drug candidates
	Future opportunities
	Conclusions
	References
Index