Machine learning and AI for healthcare : big data for improved health outcomes

목차 (Table of Contents)

• CONTENTS
• About the Author = xxi
• About the Technical Reviewer = xxiii
• Acknowledgments = xxv
• Introduction = xxvii
• Chapter 1 What Is Artificial Intelligence? = 1
• A Multifaceted Discipline = 1
• Examining Artificial Intelligence = 4
• Reactive Machines = 5
• Limited Memory : Systems That Think and Act Rationally = 5
• Theory of Mind : Systems That Think Like Humans = 5
• Self-Aware Al : Systems That Are Humans = 6
• Weak Al = 7
• Strong Al = 7
• What Is Machine Learning? = 7
• What Is Data Science? = 8
• Learning from Real-Time Big Data = 9
• Applications of Al in Healthcare = 10
• Prediction = 12
• Diagnosis = 12
• Personalized Treatment and Behavior Modification = 12
• Drug Discovery = 13
• Follow-Up Care = 13
• Realizing the Potential of Al in Healthcare = 13
• Understanding Gap = 13
• Fragmented Data = 14
• Appropriate Security = 14
• Data Governance = 15
• Bias = 15
• Software = 16
• Conclusion = 16
• Chapter 2 Data = 19
• What Is Data? = 19
• Types of Data = 21
• Big Data = 23
• Volume = 25
• Variety = 27
• Velocity = 30
• Value = 32
• Veracity = 33
• Validity = 35
• Variability = 36
• Visualization = 36
• Massive Data = 36
• Small Data = 37
• Metadata = 37
• Healthcare Data : Little and Big Use Cases = 38
• Predicting Waiting Times = 38
• Reducing Readmissions = 38
• Predictive Analytics = 38
• Electronic Health Records = 39
• Value-Based Care／Engagement = 40
• Healthcare IoT : Real-Time Notifications, Alerts, Automation = 40
• Movement Toward Evidence-Based Medicine = 42
• Public Health = 43
• Evolution of Data and Its Analytics = 43
• Turning Data into Information : Using Big Data = 45
• Descriptive Analytics = 46
• Diagnostic Analytics = 47
• Predictive Analytics = 47
• Prescriptive Analytics = 49
• Reasoning = 50
• Deduction = 50
• Induction = 51
• Abduction = 51
• How Much Data Do I Need for My Project? = 52
• Challenges of Big Data = 52
• Data Growth = 52
• Infrastructure = 52
• Expertise = 53
• Data Sources = 53
• Quality of Data = 53
• Security = 53
• Resistance = 54
• Policies and Governance = 55
• Fragmentation = 55
• Lack of Data Strategy = 55
• Visualization = 55
• Timeliness of Analysis = 56
• Ethics = 56
• Data and Information Governance = 56
• Data Stewardship = 57
• Data Quality = 57
• Data Security = 57
• Data Availability = 58
• Data Content = 58
• Master Data Management (MDM) = 58
• Use Cases = 59
• Conducting a Big Data Project = 60
• Big Data Tools = 61
• Conclusion = 62
• Chapter 3 What Is Machine Learning? = 63
• Basics = 65
• Agent = 65
• Autonomy = 65
• Interface = 66
• Performance = 66
• Goals = 66
• Utility = 66
• Knowledge = 66
• Environment = 67
• Training Data = 68
• Target Function = 68
• Hypothesis = 68
• Learner = 68
• Validation = 69
• Dataset = 69
• Feature = 69
• Feature Selection = 69
• What Is Machine Learning? = 69
• How Is Machine Learning Different from Traditional Software Engineering? = 70
• Machine Learning Basics = 72
• Supervised Learning = 72
• Unsupervised Learning = 74
• Semi-supervised = 76
• Reinforcement Learning = 76
• Data Mining = 78
• Parametric and Nonparametric Algorithms = 80
• How Machine Learning Algorithms Work = 81
• Conclusion = 83
• Chapter 4 Machine Learning Algorithms = 85
• Defining Your Machine Learning Project = 86
• Task (T) = 86
• Performance (P) = 86
• Experience (E) = 87
• Common Libraries for Machine Learning = 88
• Supervised Learning Algorithms = 90
• Selecting the Right Features = 91
• Classification = 91
• Regression = 92
• Decision Trees = 93
• Iterative Dichotomizer 3(ID3) = 96
• C4.5 = 97
• Cart = 97
• Ensembles = 98
• Bagging = 98
• Boosting = 100
• Linear Regression = 101
• Logistic Regression = 103
• SVM = 105
• Naive Bayes = 106
• kNN : k-Nearest Neighbor = 108
• Neural Networks = 109
• Perceptron = 110
• Artificial Neural Networks = 111
• Deep Learning = 113
• Feedforward Neural Network = 114
• Recurrent Neural Network (RNN) : Long Short-Term Memory = 114
• Convolutional Neural Network = 115
• Modular Neural Network = 115
• Radial Basis Neural Network = 116
• Unsupervised Learning = 117
• Clustering = 117
• k-Means = 118
• Association = 119
• Apriori = 120
• Dimensionality Reduction Algorithms = 121
• Dimension Reduction Techniques = 124
• Missing／Null Values = 124
• Low Variance = 124
• High Correlation = 124
• Random Forest Decision Trees = 124
• Backward Feature Elimination = 125
• Forward Feature Construction = 125
• Principal Component Analysis (PCA) = 125
• Natural Language Processing (NLP) = 126
• Getting Started with NLP = 128
• Preprocessing : Lexical Analysis = 128
• Noise Removal = 128
• Lexicon Normalization = 129
• Porter Stemmer = 129
• Object Standardization = 129
• Syntactic Analysis = 130
• Dependency Parsing = 130
• Part of Speech Tagging = 131
• Semantic Analysis = 132
• Techniques Used Within NLP = 132
• N-Grams = 132
• TF-IDF Vectors = 133
• Latent Semantic Analysis = 133
• Cosine Similarity = 134
• Naïve Bayesian Classifier = 135
• Genetic Algorithms = 135
• Best Practices and Considerations = 136
• Good Data Management = 136
• Establish a Performance Baseline = 137
• Spend Time Cleaning Your Data = 137
• Training Time = 138
• Choosing an Appropriate Model = 138
• Choosing Appropriate Variables = 138
• Redundancy = 138
• Overfitting = 139
• Productivity = 139
• Understandability = 140
• Accuracy = 140
• Impact of False Negatives = 140
• Linearity = 140
• Parameters = 141
• Ensembles = 141
• Use Case : Toward Smart Care in Diabetes = 141
• Predicting Blood Glucose = 142
• Predicting Risk = 142
• Predicting Risk of Other Diseases = 143
• Reversing Disease = 144
• Chapter 5 How to Perform Machine Learning = 145
• Conducting a Machine Learning Project = 145
• Framing : Specifying the Problem = 147
• Data Preparation = 149
• Training the Machine Learning Model = 150
• Evaluation and Optimization of the Method and Results = 155
• Disseminating the Results = 162
• Deployment = 165
• Conclusion = 165
• Chapter 6 Preparing Data = 167
• Phases of Data Preparation = 168
• What Is a Database? = 168
• Challenges of Databases = 169
• What Is the Difference Between a Database and a Spreadsheet? = 170
• What Is Structured Query Language (SQL)? = 171
• Common SQL Commands and Concepts = 171
• Data Gathering = 181
• Data Exploration = 181
• Unbalanced Data = 182
• Data Cleansing = 182
• Resolving Missing Values = 182
• Contradictory and Duplicate Data = 183
• Exploring Anomalies in the Data = 183
• Correcting Typos, Cleaning Values, and Formatting = 183
• Classifying Groups of Results = 184
• Transforming Data = 184
• Aggregation = 184
• Decomposition = 184
• Encoding = 185
• Scaling = 185
• Skewed Data = 185
• Bias Mitigation = 185
• Weightings = 185
• Expand = 186
• Feature Extraction = 186
• Identifying Feature Relationships = 186
• Feature Reduction = 186
• Conclusion = 187
• Chapter 7 Evaluating Machine Learning Models = 189
• Model Development and Workflow = 189
• Why Are There Two Approaches to Evaluating a Model? = 190
• Evaluation Metrics = 192
• Skewed Datasets, Anomalies, and Rare Data = 198
• Parameters and Hyperparameters = 198
• Model Parameters = 198
• Model Hyperparameters = 199
• Tuning Hyperparameters = 199
• Statistical Hypothesis Testing (Multivariate Testing) = 200
• Which Metric Should I Use for Evaluation? = 201
• Correlation Does Not Equal Causation = 202
• What Amount of Change Counts as Real Change? = 202
• Types of Tests, Statistical Power, and Effect Size = 202
• Checking the Distribution of Your Metric = 203
• Determining the Appropriate p Value = 203
• How Many Observations Are Required? = 204
• How Long to Run a Multivariate Test? = 204
• Spotting Data Variance and Drift = 204
• Keep a Note of Model Changes = 205
• Real-Time Monitoring = 205
• Conclusion = 205
• Chapter 8 Machine Learning and Al Ethics = 207
• What Is Ethics? = 209
• What Is Data Science Ethics? = 210
• Data Ethics = 210
• Informed Consent = 211
• Freedom of Choice = 212
• Should a Person's Data Consent Ever Be Overturned? = 212
• Public Understanding = 213
• Who Owns My Data? = 214
• Anonymized Data = 214
• Identifiable Data = 215
• Aggregate Data = 215
• Individualized Data = 215
• Data Controllers and Processors = 215
• What Can My Data Be Used For? = 217
• Privacy : Who Can See My Data? = 218
• Data Sharing = 218
• Anonymity Doesn't Equate to Privacy = 218
• Data Has Different Values to Different People = 219
• How Will Data Affect the Future? = 219
• Prioritizing Treatments = 219
• Determining New Treatments and Management Pathways = 219
• More Real-World Evidence = 220
• Enhancements in Pharmacology = 220
• Cybersecurity = 220
• Al and Machine Learning Ethics = 221
• What Is Machine Learning Ethics? = 221
• Machine Bias = 222
• Data Bias = 223
• Human Bias = 223
• Intelligence Bias = 223
• Bias Correction = 224
• Is Bias a Bad Thing? = 224
• Prediction Ethics = 225
• Protecting Against Mistakes = 226
• Validity = 227
• Preventing Algorithms from Becoming Immoral = 227
• Unintended Consequences = 228
• How Does Humanity Stay in Control of a Complex and Intelligent System? = 230
• What Will Happen When Al Is More Intelligent Than Humans? = 230
• Intelligence = 230
• Health Intelligence = 232
• Who Is Liable? = 233
• First-Time Problems = 234
• Defining Fairness = 235
• How Do Machines Affect Our Behavior and Interaction? = 235
• Humanity = 236
• Behavior and Addictions = 236
• Economy and Employment = 237
• Affecting the Future = 238
• Playing God = 238
• Overhype and Scaremongering = 239
• Stakeholder Buy-In and Alignment = 239
• Policy, Law, and Regulation = 239
• Data and Information Governance = 239
• Is There Such a Thing as Too Much Policy? = 240
• Global Standards and Schemas = 241
• Do We Need to Treat Al with Humanity? = 241
• Employing Data and Al Ethics Within Your Organization = 242
• Ethical Code = 242
• Ethical Framework Considerations = 244
• A Hippocratic Oath for Data Scientists = 246
• Conclusion = 247
• Chapter 9 What Is the Future of Healthcare? = 249
• Shift from Volume to Value = 250
• What Is Volume-Based Care? = 250
• Patient-Centered Care = 251
• Value-Based Care = 252
• Evidence-Based Medicine = 256
• Digital Health Research Poses a Question of Validity = 257
• How Does Evidence-Based Medicine Keep Up with the Real World? = 258
• Personalized Medicine = 258
• Personalization of Medicine Raises Ethical Issues = 259
• Personalizing Medicine with Data = 260
• Applications of Personalized Medicine = 264
• Where Else Can Al Be Used in Medicine? = 272
• Mining the EHR = 273
• Conversational Al = 273
• Making Doctors Better = 274
• Diagnosing Disease = 275
• Making and Rationalizing Decisions = 275
• Drug Discovery = 276
• 3-D Printing = 276
• Gene Therapy = 279
• Choose Your Reality = 279
• Virtual Reality = 280
• Augmented Reality = 280
• Merged Reality = 280
• Use Cases of Immersive Reality in Healthcare = 280
• Using the Blockchain in Healthcare = 283
• What Is the Blockchain? = 283
• Tamper-Proof Security = 283
• Use cases = 284
• Robots = 286
• Robot-Assisted Surgery = 286
• Exoskeletons = 287
• Inpatient Care = 287
• Companions = 287
• Drones = 287
• Smart Places = 288
• Smart Homes = 289
• Smart Hospitals = 289
• Developing Whole Al = 290
• Conclusion = 291
• Chapter 10 Case Studies = 293
• Real-World Inspiration = 293
• Real-World Application and Learnings = 294
• Case Study 1 : Al for Imaging of Diabetic Foot Concerns and Prioritization of Referral for Improvements in Morbidity and Mortality = 295
• Background = 296
• Cognitive Vision = 297
• Project Aims = 298
• Challenges = 299
• Conclusions = 301
• Case Study 2 : Outcomes of a Digitally Delivered, Low-Carbohydrate, Type 2 Diabetes Self-Management Program: 1-Year Results of a Single-Arm Longitudinal Study = 302
• Background = 302
• Objectives = 303
• Methods = 303
• Results = 304
• Observations = 305
• Conclusions = 306
• Case Study 3 : Delivering a Scalable and Engaging Digital Therapy for Epilepsy = 306
• Background = 306
• Implementing the Evidence Base = 306
• Sensor-Driven Digital Program = 307
• Research = 308
• Project Impact = 308
• Preliminary Analysis = 309
• Case Study 4 : Improving Learning Outcomes for Junior Doctors Through the Novel Use of Augmented and Virtual Reality = 309
• Background = 310
• Aims = 310
• Project Description = 311
• What Is 360° Video? = 311
• Conclusions = 312
• Case Study 5 : Do Wearable Apps Have Any Effect on Health Outcomes? A Real-World Service Evaluation of the Impact on Activity = 313
• Background = 313
• Methods = 315
• Results = 318
• Discussion = 320
• Conclusions = 321
• Case Study 6 : Big Data, Big Impact, Big Ethics : Diagnosing Disease Risk from Patient Data = 321
• Background = 321
• Platform Services = 322
• Medication Adherence, Efficacy, and Burden = 322
• Big Data : Pooling People to Empower Health Decisions = 323
• Al Prioritization of Patient Interactions = 324
• Real-World Evidence = 325
• Ethical Implications of Predictive Analytics = 326
• Integration of the loT = 327
• Conclusions = 327
• Case Study 7 : Assessment of a Predictive Al Model for Personalised Care and Evaluation of Accuracy = 328
• Background = 328
• Problem Statement = 329
• Expectation-Based Modeling Approach = 329
• Model Evaluation = 333
• Improving the Solution = 335
• Conclusion = 337
• Case Study 8 : Can Voice-Activated Assistants Support Adults to Remain Autonomous, a Real-World Service Evaluation of the Impact of a Voice-Activated Smart Speaker Application on Weight and Activity = 338
• Background = 338
• Objective = 342
• Methods = 342
• Results = 345
• Discussion = 347
• Conclusion = 349
• Appendix A : References = 351
• Appendix B : Technical Glossary = 385
• Index = 399