Artificial intelligence (AI) and machine learning (ML) techniques play an important role in our daily lives by enhancing predictions and decision-making for the public in several fields such as financial services, real estate business, consumer goods, social media, etc. Despite several studies that have proved the efficacy of AI/ML tools in providing improved healthcare solutions, it has not gained the trust of health-care practitioners and medical scientists. This is due to poor reporting of the technology, variability in medical data, small datasets, and lack of standard guidelines for application of AI. Therefore, the development of new AI/ML tools for various domains of medicine is an ongoing field of research.
Machine Learning in Healthcare: Fundamentals and Recent Applications discusses how to build various ML algorithms and how they can be applied to improve healthcare systems. Healthcare applications of AI are innumerable: medical data analysis, early detection and diagnosis of disease, providing objective-based evidence to reduce human errors, curtailing inter- and intra-observer errors, risk identification and interventions for healthcare management, real-time health monitoring, assisting clinicians and patients for selecting appropriate medications, and evaluating drug responses. Extensive demonstrations and discussion on the various principles of machine learning and its application in healthcare is provided, along with solved examples and exercises.
This text is ideal for readers interested in machine learning without any background knowledge and looking to implement machine-learning models for healthcare systems.
Author(s): Bikesh Kumar Singh, G.R. Sinha
Publisher: CRC Press
Year: 2022
Language: English
Pages: 232
City: Boca Raton
Cover
Half Title
Title Page
Copyright Page
Dedication
Table of Contents
Figures
Tables
Preface
Acknowledgments
Author Bio
1 Biostatistics
Learning Objectives
1.1 Data and Variables
1.2 Types of Research Studies
1.3 Sources of Medical Data
1.4 Measures of Central Tendency
1.5 Data Sampling and Its Types
1.5.1 Probability Sampling Methods
1.5.2 Non-Probability Sampling Methods
1.6 Statistical Significance Analysis
1.7 Skewness
1.8 Kurtosis
1.8.1 Mesokurtic
1.8.2 Leptokurtic
1.8.3 Platykurtic
1.9 Curve Fitting
1.9.1 Linear and Non-Linear Relationship
1.9.2 Use of Curve-Fitting Method
1.10 Correlation
1.10.1 Pearson Correlation (PC)
1.10.2 Spearman Rank Correlation (SRC)
1.11 Regression
1.11.1 Linear Regression
1.11.2 Estimation of Regression Coefficients
1.12 Learning Outcomes
2 Probability Theory
Learning Objectives
2.1 Basic Concept of Probability
2.2 Random Experiment
2.3 Conditional Probability
2.3.1 Types of Events
2.4 Bayes Theorem
2.5 Random Variable
2.6 Distribution Functions
2.6.1 Binomial Distribution
2.6.2 Poisson Distribution
2.6.3 Normal Distribution
2.7 Estimation
2.8 Standard Error
2.9 Probability of Error
Learning Outcomes
3 Medical Data Acquisition and Pre-Processing
Learning Objectives
3.1 Medical Data Formats
3.1.1 Data Formats for Medical Images
3.1.1.1 DICOM (Digital Imaging and Communications in Medicine)
3.1.1.2 Analyse
3.1.1.3 NIfTI (Neuroimaging Informatics Technology Initiative)
3.1.1.4 MINC (Medical Imaging NetCDF)
3.1.2 Medical Data Formats for Signals
3.1.2.1 EDF (European Data Format)
3.1.2.2 BDF (BioSemi Data Format)
3.1.2.3 GDF (General Data Format)
3.2 Data Augmentation and Generation
3.3 Data Labelling
3.4 Data Cleaning
3.4.1 Statistical Approach
3.4.1.1 Listwise Deletion
3.4.1.2 Pairwise Deletion
3.4.1.3 Multiple Imputation
3.4.1.4 Maximum Likelihood Imputation
3.4.2 Machine Learning for Data Imputation
3.4.2.1 K-Nearest Neighbour (KNN)
3.4.2.2 Bayesian Network (BN)
3.5 Data Normalization
Learning Outcomes
References
4 Medical Image Processing
Learning Objectives
4.1 Medical Image Modalities, Their Applications, Advantages and Limitations
4.1.1 Radiography
4.1.2 Nuclear Medicine
4.1.2.1 Positron Emission Tomography (PET)
4.1.3 Elastography
4.1.4 Photoacoustic Imaging
4.1.5 Tomography
X-ray Computed Tomography (CT)
4.1.6 Magnetic Resonance Imaging (MRI)
4.1.7 Ultrasound Imaging Techniques
4.2 Medical Image Enhancement
4.3 Basics of Histogram
4.4 Medical Image De-Noising
4.4.1 Spatial Filtering
4.4.1.1 Linear Filters
4.4.1.2 Non-Linear Filters
4.4.2 Transform Domain Filtering
4.4.2.1 Non-Data Adaptive Transform
4.4.2.2 Data-Adaptive Transforms
4.5 Segmentation
4.6 Region-Based Methods
4.6.1 Region-Growing Segmentation
Following Are Advantages and Disadvantages of the Region-Growing Method of Segmentation
Advantages
Disadvantages
Learning Outcomes
References
5 Bio-Signals
Learning Objectives
5.1 Origin of Bio-Signals
5.2 Different Types of Bio-Signals
5.2.1 Electrocardiogram
5.2.2 Electroencephalogram (EEG)
5.2.3 Electroocculogram (EOG)
5.2.4 Electromyogram (EMG)
5.3 Noise and Artefacts
5.4 Filtering of Bio-Signals
5.5 Applications of Bio-Signals
Learning Outcome
References
6 Feature Extraction
Learning Objectives
6.1 Feature Extraction
6.2 Echographic Characteristics of Breast Tumours in Ultrasound Imaging
6.3 Texture Feature Extraction
6.3.1 First-Order Statistical Features
6.3.2 Grey-Level Co-Occurrence Matrices
6.3.3 Grey-Level Difference Statistics
6.3.4 Neighbourhood Grey-Tone Difference Matrix
6.3.5 Statistical Feature Matrix
6.3.6 Texture Energy Measures
6.3.7 Fractal Dimension Texture Analysis
6.3.8 Spectral Measures of Texture
6.3.9 Run-Length Texture Features
6.4 Shape Feature Extraction
6.4.1 Region Properties
6.4.2 Moment Invariants
6.5 Feature Normalization
6.5.1 Brief Overview of Feature Normalization Techniques
Learning Outcome
References
7 Introduction to Machine Learning
Learning Objectives
7.1 Introduction: What Is Machine Learning?
Some Terminology of Machine Learning
7.2 Classification of Machine Learning (ML) Methods
7.3 Steps in Implementation of Machine Learning
Step 1: Gathering Data
Step 2: Preparing the Data
Step 3: Choosing a Model
Step 4: Training
Step 5: Evaluation
Step 6: Hyperparameter Tuning
Step 7: Prediction
7.4 Training, Testing and Validation
7.5 Machine Learning Methods
7.5.1 Supervised Learning
Naive Bayes Classifier
Support Vector Machine (SVM)
K-Nearest Neighbour Classifier
Decision Tree Classifier
Random Forest
Artificial Neural Network (ANN)
7.5.2 Unsupervised Learning
K-Means Clustering
Fuzzy C-Means Clustering
7.6 Performance Evaluation of Machine Learning Model
Learning Outcomes
Exercise
References
8 Cancer Detection: Breast Cancer Detection Using Mammography, Ultrasound and Magnetic Resonance Imaging (MRI)
Learning Objectives
8.1 Introduction
8.2 Different Imaging Modalities
8.2.1 Mammography (MG)
8.2.2 Ultrasound (US)
8.2.3 Magnetic Resonance Imaging (MRI)
8.3 Breast Imaging Reporting and Data System (BI-RADS)
8.4 Usefulness of Machine Learning (ML)
8.4.1 Image Pre-Processing
8.4.2 Image Segmentation
8.4.3 Feature Extraction
8.4.4 Feature Selection
8.4.5 Classification
8.4.6 Performance Evaluation
8.5 Issues and Challenges
8.6 Conclusion
Learning Outcomes
Exercise for Practice and Discussion
References
9 Sickle Cell Disease Management: A Machine Learning Approach
Learning Objectives
9.1 Introduction
9.2 Severity Detection of Sickle Cell Disease
9.2.1 Analysis of Clinical Complications
9.2.2 Analysis of Clinical Attributes
9.2.3 Analysis of Microscopic Images of RBC
9.3 Hydroxyurea Dosage Prediction for SCD Patients
9.4 Patient Response to Medications Through Hydroxyurea (HU)
9.5 SCD Management Proposed Model
9.6 Conclusions
Learning Outcomes
References
10 Detection of Pulmonary Disease
Learning Objectives
10.1 Introduction to Pulmonary Disorders
10.2 Restrictive and Obstructive Lung Diseases
10.2.1 Obstructive Lung Disease
10.2.2 Restrictive Lung Disease
10.3 Diagnosis of Disease and Disorder
10.4 Chest X-Ray
10.5 CT Scan
10.6 SPO2 Level
10.7 Arterial Blood Gas Analysis
10.8 Laboratory Tests
10.9 Bronchoscopy
10.10 Sputum Test
10.11 Pulmonary Function Test
10.12 Challenges and Issues
10.13 Application of Machine Learning in Diagnosis of Pulmonary Disorder
10.14 Conclusion
Learning Outcomes
Exercise for Practice and Discussion
References
11 Mental Illness and Neurodevelopmental Disorders
Learning Objectives
11.1 Neurodevelopmental Disorders
11.2 Developmental Dyslexia
11.2.1 Diagnostic Methods
11.2.2 Behavioural Method
11.2.3 Brain Imaging Modalities
11.2.4 Recent Advancement in Diagnostic Techniques
11.3 Attention-Deficit/Hyperactivity Disorder (ADHD)
11.3.1 Types
11.3.2 Symptoms
11.3.3 ADHD Screening
11.3.4 Diagnosis Based On Brain Imaging and Machine Learning Methods
11.3.5 Treatment for ADHD
11.4 Parkinson’s Disease
11.4.1 Parkinson’s Disease Prognosis and Measurement Rating Scales
11.4.1.1 HY Scale
11.4.1.2 UPDRS Scale
11.4.2 Involvement of Digital Technologies for Detection and Monitoring of PD
11.5 Epilepsy
11.5.1 Recent Literatures On Epilepsy Detection
11.5.2 Generalized Machine Learning Model for Epilepsy Detection System
11.6 Schizophrenia
11.6.1 Recent Research
11.6.2 A Machine Learning Model for Schizophrenia Detection
Learning Outcomes
References
12 Applications and Challenges
Learning Objectives
12.1 Role of Machine Learning in Healthcare Research
12.2 Efficient Diagnosis of Diabetes
12.3 Neuropathy
12.4 Drug Monitoring
12.5 Bioinformatics
12.6 DNA Analysis
12.7 Digital Health Records
12.8 Future Research Challenges
Learning Outcomes
References
Index
