Bank Fraud (eBook)

Revathi Subramanian is Senior Vice President, Data Science at CA Technologies, which helps Fortune 1000 companies manage and secure complex IT environments to support agile business services. She is the founding member of a team of high caliber data scientists that are uncovering business value and operational intelligence from the chaos of Big Data in areas like eCommerce, application performance management, infrastructure management, service virtualization, and project management. Before joining CA, Revathi was the co-founder of the SAS Advanced Analytic Solutions Division in 2002. She led the development of a new enterprise real-time fraud decisioning platform utilizing advanced analytics. Revathi has a Master's degree in Statistics from The Ohio State University and a Bachelor's degree in Mathematics from Ethiraj Collge, Chennai, India.

Preface xi

Acknowledgments xiii

About the Author xvii

Chapter 1 Bank Fraud: Then and Now 1

The Evolution of Fraud 2

The Evolution of Fraud Analysis 8

Summary 14

Chapter 2 Quantifying Fraud: Whose Loss Is It Anyway?
15

Fraud in the Credit Card Industry 22

The Advent of Behavioral Models 30

Fraud Management: An Evolving Challenge 31

Fraud Detection across Domains 33

Using Fraud Detection Effectively 35

Summary 37

Chapter 3 In God We Trust. The Rest Bring Data! 39

Data Analysis and Causal Relationships 40

Behavioral Modeling in Financial Institutions 42

Setting Up a Data Environment 47

Understanding Text Data 58

Summary 60

Chapter 4 Tackling Fraud: The Ten Commandments 63

1. Data: Garbage In; Garbage Out 67

2. No Documentation? No Change! 71

3. Key Employees Are Not a Substitute for Good Documentation
75

4. Rules: More Doesn't Mean Better 77

5. Score: Never Rest on Your Laurels 79

6. Score + Rules = Winning Strategy 83

7. Fraud: It Is Everyone's Problem 85

8. Continual Assessment Is the Key 86

9. Fraud Control Systems: If They Rest, They Rust 87

10. Continual Improvement: The Cycle Never Ends 88

Summary 88

Chapter 5 It Is Not Real Progress Until It Is Operational
89

The Importance of Presenting a Solid Picture 90

Building an Effective Model 92

Summary 105

Chapter 6 The Chain Is Only as Strong as Its Weakest Link
109

Distinct Stages of a Data-Driven Fraud Management System 110

The Essentials of Building a Good Fraud Model 112

A Good Fraud Management System Begins with the Right Attitude
117

Summary 119

Chapter 7 Fraud Analytics: We Are Just Scratching the Surface
121

A Note about the Data 125

Data 126

Regression 1 128

Logistic Regression 1 132

"Models Should Be as Simple as Possible, But Not
Simpler" 149

Summary 151

Chapter 8 The Proof of the Pudding May Not Be in the Eating
153

Understanding Production Fraud Model Performance 154

The Science of Quality Control 155

False Positive Ratios 156

Measurement of Fraud Detection against Account False Positive
Ratio 156

Unsupervised and Semisupervised Modeling Methodologies 158

Summary 159

Chapter 9 The End: It Is Really the Beginning! 161

Notes 165

Index 167