Machine Learning and Artificial Intelligence in Healthcare Systems: Tools and Techniques

Cover
Half Title
Series Page
Title Page
Copyright Page
Contents
1. Artificial Intelligence Challenges, Principles, and Applications in Smart Healthcare Systems
Introduction
The Smart Healthcare Concept
Primary Objectives for AI Applications
Artificial Intelligence Benefits in Health Care
Healthcare Benefits of Artificial Intelligence
Accessibility Improved
Early Diagnosis
Speed Upgrade and Cost Savings
Surgery Assistance that is Effective and Unique
Support for Mental Health and Enhanced Human Ability
Artificial Intelligence Challenges in Healthcare
Digitization and Consolidation of Data
Updating Regulations
Human Interventions
Prospective V / s Retrospective Research
As an Evidence Gold Standard, Randomized Controlled Trials with Peer Review
Clinical Value is not Always Reflected in Metrics
Comparing Different Algorithms is Difficult
Difficulties in the Science of Machine Learning
Shift in Dataset
Fitting Confounders by Accident vs. True Signal
Difficulties in Generalisation to New Populations and Settings
Bias in the Algorithmic Process Bias
Susceptibility to Adversarial Attack or Manipulation
Difficulties in Logistically Implementing AI Systems
Achieving Robust Regulation and Rigorous Quality Control
Human Resistance to Artificial Intelligence (AI) in Healthcare
Improving our Understanding of the Humans' and Algorithms' Interaction
Risks of Artificial Intelligence in Healthcare
Injuries and Errors
Privacy Issues
Discrimination and Inequality
Professional Reshuffling
In the Healthcare System, Applications of Artificial Intelligence (AI)
Application of Artificial Intelligence in Modern Medicine
Alginate and Artificial Intelligence in Biomedical Fields
Conclusion
References
2. Systematic View and Impact of Artificial Intelligence in Smart Healthcare Systems, Principles, Challenges and Applications
Introduction
Artificial Intelligence: A Reference Point for Innovation
Practice and Design Principles
Artificial Intelligence's growth in Healthcare Domain
Professional Support to Health Care in AI
Performance Indicators for the progress of Health Care in AI
Applications in Diagnosis and Treatment
Inference about Healthcare
Fear and Expectations about AI
Expectations from AI in Healthcare
More Efficient Healthcare Logistics
Taking Drug Development to New Levels
Improve the Working Conditions of Medical Workers While Saving Lives
Identifying Novel Links Between Risks and Illnesses
Bringing the Art of Medicine Into a New Age
Assist in the Prediction of Future Outbreaks and Pandemics
Artificial Intelligence Will Not Assist You in Healthcare
Medical Practitioners will be Replaced by Artificial Intelligence
Artificial Intelligence will Have the same Level of Understanding as a Real-Life Doctor Patient Interaction
In Healthcare, Artificial Intelligence will Help with Privacy Difficulties
Artificial Intelligence will Develop Completely Autonomous Surgical Robots
Medical Choices will be Made Solely by Artificial Intelligence
AI Will not be Prejudiced
AI Will Think in the Same way that Humans Do
Overview of current AI in Healthcare
Radiology for Diagnosis
Pathology
Ophthalmology
Cardiology
Challenges and Solutions of AI in healthcare
Artificial Intelligence Model Development and Validation
Model Development
Important Considerations in Model Development
Learning in Model Development
Model Validation Framework
AI/ML Solution Hazard Tiering
Fairness & Bias
Validating the Model
Healthcare tools using AI
Amazon's Alexa Voice Assistant has a New Trick
Health Chatbots
OneRemission
Youper
Babylon Health
Florence
Healthily
Ada Health
Sensely
Buoy Health
Infermedica
GYANT
Woebot
Cancer Chatbot
Case Study
Using IBM Watson to aid Oncologists in India is a Case Study
The Influence of AI on Data Privacy and Ethical Issues
Examining the Impact of Artificial Intelligence on Physicians
Conclusion
References
3. Application of Machine Learning Techniques in COVID-19 Epidemiology: A Glimpse
Introduction
Attributes of COVID-19 Epidemiology
Distribution
Geographic Distribution
Disease Trajectory
Determinants
Transmission Dynamics
Clinical Features
Immune Response
Risk of Reinfection
Machine Learning in COVID-19 Epidemiology at a Glance
Distribution
Geographic Distribution
Disease Trajectory
Determinants
Transmission Dynamics of COVID-19
Clinical Features
Immune Response
Risk of Reinfection
Conclusion
References
4. Automated Seven-Level Skin Cancer Staging Diagnosis in Dermoscopic images using Deep Learning
Introduction
Skin Cancer
Stages of Skin Cancer
Limitation of Machine Learning
Background and Related Work
Theoretical Concepts
Preprocessing in Deep Learning
Normalization
Standardization
Classification
Convolutional Neural Network (CNN)
Layers in Convolutional Neural Network
Filter (or) Kernel
Padding
Strided Convolution
Convolution over Volume
Pooling Layer
Activation Function
Soft Max
Convolutional Layer
Fully-connected Layers
Methods and Results
Proposed Methodology
Dataset
Preprocessing
Splitting the Dataset
Classification
Performance Analysis
Expected Result and Predicted Result
Confusion Matrix
Conclusion and Future Enhancement
References
5. Ensemble Classifier Based Predictive Model for Type-2 Diabetes Mellitus Prediction
Introduction
Literature Survey
Theoretical Concepts Used
Boosting
Gradient Boosting Classifier
Light Gradient Boosting Machine (LGBM)
Bagging
Random Forest
Proposed System
Methodology
Dataset Collection and Description
Preprocessing of Data
Exploratory Data Analysis (EDA)
Feature Correlation
Feature Engineering
Model Development Process
Model Ensemble
Hyper Parameter Tuning
Model Evaluation
Receiver Operating Characteristics (ROC)
Results and Discussion
Conclusion and Future Work
References
6. Machine Learning Approaches for Analysis in Smart Healthcare Informatics
Introduction
Smart Health
Smart Healthcare Applications
Diagnosis
Healthcare Administration
Diseases and Risk Analysis
Smart Hospitals that are Forward-Thinking
Assisting with Drug Development
IoT and Healthcare
Machine Learning Algorithms
Learning under Supervision
Learning Without Supervision
Hyperparameters
Reinforcement Learning in Health Care
Applications of Reinforcement Learning in Healthcare
Dynamic Treatment Regimens (DTRs)
Medical Diagnostic
Scheduling and Allocating Healthcare Resources
Drug Research, Development, and Design
Health Administration
The Difficulties of Reinforcement Learning in Healthcare
Scarcity of Data
Partially Observable
Formulation and Design of Rewards
Case Study
An Extensive Medicare Data Exploration
CMS (Content Management System)
Conclusion and Future Scope
References
7. Smart Approaches for Diagnosis of Brain Disorders Using Artificial Intelligence
Introduction
Brain MRI
Physics of MRI
MRI Imaging Sequences
Application
Brain MRI Artifacts
Use of Deep Learning in Health Diagnosis
Deep Learning in Medical Imaging
Deep Learning Methods
Deep Learning in the Brain
Electroencephalogram (EEG)
Characteristic Nature of Electroencephalogram (EEG) Signals
EEG Signal Analysis and Classification
EEG Data Processing
Preprocessing
Feature Extraction
Classification
Deep Learning for Detection of Brain Disorders using EEG
Conclusion
References
8. Bridging the Gap Between Technology and Medicine: Approaches of Artificial Intelligence in Healthcare
Introduction
Application Areas of Artificial Intelligence in Healthcare
Diagnosis and Treatment of Diseases
Electronic Health Records (EHRs)
Drug Discovery
Radiology
Machine Learning Algorithms in Healthcare Industry
Naive Bayes
Logistic Regression
k-Nearest Neighbor
Support Vector Machines
Deep Learning in Healthcare
Performance Metrics for Model Evaluation
Confusion Matrix
True Positives (TP)
True Negatives (TN)
False Positives (FP)
False Negatives (FN)
Precision
Recall
F1-score
Accuracy
AUC-ROC Curve
Logarithmic Loss (Log Loss)
Challenges to Artificial Intelligence in Healthcare
Informed Consent
Safety, Transparency, and Effectiveness
Fairness and bias of algorithms
Data privacy
Conclusion
References
9. Brain Tumor Classification Using Transfer Learning
Introduction
Related Works
Convolutional Neural Network (CNN) based Approach
Support Vector Machine (SVM) based Approach
Challenges
Proposed Model
Data PreProcessing
Canny Edge Detection
Data Augmentation
Transfer Learning
EfficientNet B0
ResNet50
DenseNet121
MobileNetV3-Small
Feature Fusion
Experimental Framework
Implementation Details
Loss Function
Optimization Function
Dataset Description
Results and Discussion
EfficientNet B0
ResNet50
DenseNet121
MobileNetV3-Small
Conclusion and Future Works
References
10. Advanced Bayesian Estimation of Weibull in Early Stage Eye Loss Prediction in Diabetic Retinopathy
Introduction
Literature Review
Bayesian Model and Survival Model
Bayesian Model
Naïve Bayes
Tree-Augmented Naive Bayes (TAN)
Survival Model
Censored Data
Survival Function
Proposed Approach: Bayesian Estimation of Weibull in Early Stage
Formulation Approach
Process Flow Diagram
Prior Probability Extrapolation
Implementation and Results
Dataset Description
Performance Evaluation
Results and Discussion
AUC Measure
Accuracy Measure
F-Measure
Conclusion and Future Scope
References
11. Automated Sleep Staging Using Single-Channel EEG Signal Based on Machine Learning Approaches
Introduction
Motivation
Importance of Human Sleep Study
Sleep Structure and Sleep Stages
Sleep Stages Behavior
Experimental Studies on Sleep Staging System
Experimental Data
Proposed Automatic Sleep Stage Detection Method
Experimental Results and Discussion
Classification Accuracy of Category-I Subject ISRUC-Sleep Database
Classification Accuracy of Category-II Subject ISRUC-Sleep Database
Classification Accuracy of Category-III Subject ISRUC-Sleep Database
Summary of Comparative Analysis Results of ISRUC-Sleep (SG-I/SG-II/SG-III) data
Conclusion
References
12. Machine Learning-Based Intelligent Assistant for Smart Healthcare
Introduction
Literature Review
Motivation for the Proposed Work
Application of AI/Ml in Healthcare Industry
Application of Chatbots in Healthcare
Robots in Medical Labs
Medical Image Diagnostics with AI and ML in Healthcare
AI Empowered Health Companions
Proposed ML Equipped Intelligent Solution For Smart Healthcare
Application of ML Algorithms in Healthcare Prediction
Logistic Regression
K-Nearest Neighbor Algorithm (KNN Algorithm)
Support Vector Machine (SVM)
Gaussian Naive Bayes
Decision Tree
Random Forest
Gradient Boost
Logistic Regression ML Algorithm
Case Study - Patient Data Analytics Using Logistic Regression ML Algorithm in Intelligent Virtual Assistant
Application of Logistic Regression Model in Prediction Analysis
Problem to be Addressed: Stroke Prediction
Predicting Stroke Using Datasets
Dataset Taken for Learning Purpose
Attributes Information
Unique Values for Attributes
Observations
Prediction Variable (Desired Target)
Sample Python Code for Logistic Regression Model to Predict Stroke Disease
Data Preprocessing
Observations
Statistical Analysis to assess the Prediction Efficiency
Model Training and Prediction using Logarithmic Regression Model
Summary of Findings from the Analysis
Conclusion and Future Research Directions
Conclusion
Future Research Directions
References
Additional Readings
13. AI-enabled Sentiment Analysis on COVID-19 Vaccination: A Twitter Based Study
Introduction
Background and Literature Review
Word2vec
N-gram
Tf-Idf
VADER (Valence Aware Dictionary and sEntiment Reasoner)
Pattern
Sentiment Analysis of Twitter Data Related to Vaccines: A Case Study
Dataset Description
Data Preprocessing
Exploratory Data Analysis
Overall Sentiment Analysis Through VADER Tool
Overall Sentiment Analysis Through Pattern Library
Analysis of AstraZeneca Related Tweets
Analysis of Pfizer Related Tweets
Analysis of Covaxin Related Tweets
Analysis of Moderna Related Tweets
Analysis of Sputnik V Related Tweets
Conclusion
Notes
References
14. An Early Diagnosis of Lung Nodule Using CT Images Based on Hybrid Machine Learning Techniques
Introduction
Related Works
Problem Methodology
Research Gap
System Model
Proposed EPC-HML Techniques
Lung Nodule Segmentation Using ACS-ML
Feature Extraction and Selection Using HBS Algorithm
Lung Nodule Classification Using SM-CNN
Simulation Results and Discussion
Description of the Datasets
Training and Implementation
Advantages and Limitations of Lung Nodule Diagnosis
Analysis of Segmentation and Feature Extraction Process
Comparative Analysis of Classifiers
Conclusion and Future Work
References
15. Early Detection of Alzheimer's Disease Assisted by AI-Powered Human-Robot Communication
Overview of Application of AI-powered Robots for Early Detection of Alzheimer's Disease (AD) Using Human-Robot Communication
Introduction
Alzheimer's Disease (AD)
Detection of Alzheimer's Disease
Human-robot Communication
Human-robot Communication in Early Detection of AD
Early Detection of AD Using Human-Robot Communication: A Scoping Review
Results
Description of Robots
Performance Evaluation and Outcomes
AI-powered Human-Robot Communication for Advancement of Early Detection of AD in Healthcare Services
Study Participant Feedback
The Future Use of Human-Robot Intervention for Early Detection of AD
Conclusions
Acknowledgment
References
Index