Bart Baesens 
Analytics in a Big Data World [PDF ebook] 
The Essential Guide to Data Science and its Applications

Preface xiii
Acknowledgments xv
Chapter 1 Big Data and Analytics 1
Example Applications 2
Basic Nomenclature 4
Analytics Process Model 4
Job Profiles Involved 6
Analytics 7
Analytical Model Requirements 9
Notes 10
Chapter 2 Data Collection, Sampling, and Preprocessing 13
Types of Data Sources 13
Sampling 15
Types of Data Elements 17
Visual Data Exploration and Exploratory Statistical Analysis 17
Missing Values 19
Outlier Detection and Treatment 20
Standardizing Data 24
Categorization 24
Weights of Evidence Coding 28
Variable Selection 29
Segmentation 32
Notes 33
Chapter 3 Predictive Analytics 35
Target Definition 35
Linear Regression 38
Logistic Regression 39
Decision Trees 42
Neural Networks 48
Support Vector Machines 58
Ensemble Methods 64
Multiclass Classification Techniques 67
Evaluating Predictive Models 71
Notes 84
Chapter 4 Descriptive Analytics 87
Association Rules 87
Sequence Rules 94
Segmentation 95
Notes 104
Chapter 5 Survival Analysis 105
Survival Analysis Measurements 106
Kaplan Meier Analysis 109
Parametric Survival Analysis 111
Proportional Hazards Regression 114
Extensions of Survival Analysis Models 116
Evaluating Survival Analysis Models 117
Notes 117
Chapter 6 Social Network Analytics 119
Social Network Definitions 119
Social Network Metrics 121
Social Network Learning 123
Relational Neighbor Classifier 124
Probabilistic Relational Neighbor Classifier 125
Relational Logistic Regression 126
Collective Inferencing 128
Egonets 129
Bigraphs 130
Notes 132
Chapter 7 Analytics: Putting It All to Work 133
Backtesting Analytical Models 134
Benchmarking 146
Data Quality 149
Software 153
Privacy 155
Model Design and Documentation 158
Corporate Governance 159
Notes 159
Chapter 8 Example Applications 161
Credit Risk Modeling 161
Fraud Detection 165
Net Lift Response Modeling 168
Churn Prediction 172
Recommender Systems 176
Web Analytics 185
Social Media Analytics 195
Business Process Analytics 204
Notes 220
About the Author 223
Index 225

BART BAESENS is an associate professor at KU Leuven (Belgium) and a lecturer at the University of Southampton (United Kingdom), as well as an internationally known data analytics consultant. He is a foremost researcher in the areas of web analytics, customer relationship management, and fraud detection. His findings have been published in well-known international journals including Machine Learning and Management Science. Baesens is also co-author of the book Credit Risk Management: Basic Concepts (Oxford University Press, 2008).