Fraud Analytics Using Descriptive, Predictive, and Social Network Techniques

A Guide to Data Science for Fraud Detection

Bart Baesens, Veronique Van Vlasselaer, Wouter Verbeke

• Éditeur: Wiley
• Année: 2015
• ISBN:
  9781119146834
• Édition: 1
• Sujet: Computer Science
• Sous-sujet: Data Mining

Detect fraud earlier to mitigate loss and prevent cascading damage

Fraud Analytics Using Descriptive, Predictive, and Social Network Techniques is an authoritative guidebook for setting up a comprehensive fraud detection analytics solution. Early detection is a key factor in mitigating fraud damage, but it involves more specialized techniques than detecting fraud at the more advanced stages. This invaluable guide details both the theory and technical aspects of these techniques, and provides expert insight into streamlining implementation. Coverage includes data gathering, preprocessing, model building, and post-implementation, with comprehensive guidance on various learning techniques and the data types utilized by each. These techniques are effective for fraud detection across industry boundaries, including applications in insurance fraud, credit card fraud, anti-money laundering, healthcare fraud, telecommunications fraud, click fraud, tax evasion, and more, giving you a highly practical framework for fraud prevention.

It is estimated that a typical organization loses about 5% of its revenue to fraud every year. More effective fraud detection is possible, and this book describes the various analytical techniques your organization must implement to put a stop to the revenue leak.

• Examine fraud patterns in historical data
• Utilize labeled, unlabeled, and networked data
• Detect fraud before the damage cascades
• Reduce losses, increase recovery, and tighten security

The longer fraud is allowed to go on, the more harm it causes. It expands exponentially, sending ripples of damage throughout the organization, and becomes more and more complex to track, stop, and reverse. Fraud prevention relies on early and effective fraud detection, enabled by the techniques discussed here. Fraud Analytics Using Descriptive, Predictive, and Social Network Techniques helps you stop fraud in its tracks, and eliminate the opportunities for future occurrence.

Chapter 1
Fraud: Detection, Prevention, and Analytics!

Introduction

In this first chapter, we set the scene for what's ahead by introducing fraud analytics using descriptive, predictive, and social network techniques. We start off by defining and characterizing fraud and discuss different types of fraud. Next, fraud detection and prevention is discussed as a means to address and limit the amount and overall impact of fraud. Big data and analytics provide powerful tools that may improve an organization's fraud detection system. We discuss in detail how and why these tools complement traditional expert-based fraud-detection approaches. Subsequently, the fraud analytics process model is introduced, providing a high-level overview of the steps that are followed in developing and implementing a data-driven fraud-detection system. The chapter concludes by discussing the characteristics and skills of a good fraud data scientist, followed by a scientific perspective on the topic.

Fraud!

Since a thorough discussion or investigation requires clear and precise definitions of the subject of interest, this first section starts by defining fraud and by highlighting a number of essential characteristics. Subsequently, an explanatory conceptual model will be introduced that provides deeper insight in the underlying drivers of fraudsters, the individuals committing fraud. Insight in the field of application—or in other words, expert knowledge—is crucial for analytics to be successfully applied in any setting, and matters eventually as much as technical skill. Expert knowledge or insight in the problem at hand helps an analyst in gathering and processing the right information in the right manner, and to customize data allowing analytical techniques to perform as well as possible in detecting fraud.

The Oxford Dictionary defines fraud as follows:

---

Wrongful or criminal deception intended to result in financial or personal gain.

---

On the one hand, this definition captures the essence of fraud and covers the many different forms and types of fraud that will be discussed in this book. On the other hand, it does not very precisely describe the nature and characteristics of fraud, and as such, does not provide much direction for discussing the requirements of a fraud detection system. A more useful definition will be provided below.

Fraud is definitely not a recent phenomenon unique to modern society, nor is it even unique to mankind. Animal species also engage in what could be called fraudulent activities, although maybe we should classify the behavior as displayed by, for instance, chameleons, stick insects, apes, and others rather as manipulative behavior instead of fraudulent activities, since wrongful or criminal are human categories or concepts that do not straightforwardly apply to animals. Indeed, whether activities are wrongful or criminal depends on the applicable rules or legislation, which defines explicitly and formally these categories that are required in order to be able to classify behavior as being fraudulent.

A more thorough and detailed characterization of the multifaceted phenomenon of fraud is provided by Van Vlasselaer et al. (2015):

---

Fraud is an uncommon, well-considered, imperceptibly concealed, time-evolving and often carefully organized crime which appears in many types of forms.

---

This definition highlights five characteristics that are associated with particular challenges related to developing a fraud-detection system, which is the main topic of this book. The first emphasized characteristic and associated challenge concerns the fact that fraud is uncommon. Independent of the exact setting or application, only a minority of the involved population of cases typically concerns fraud, of which furthermore only a limited number will be known to concern fraud. This makes it difficult to both detect fraud, since the fraudulent cases are covered by the nonfraudulent ones, as well as to learn from historical cases to build a powerful fraud-detection system since only few examples are available.

In fact, fraudsters exactly try to blend in and not to behave different from others in order not to get noticed and to remain covered by nonfraudsters. This effectively makes fraud imperceptibly concealed, since fraudsters do succeed in hiding by well considering and planning how to precisely commit fraud. Their behavior is definitely not impulsive and unplanned, since if it were, detection would be far easier.

They also adapt and refine their methods, which they need to do in order to remain undetected. Fraud-detection systems improve and learn by example. Therefore, the techniques and tricks fraudsters adopt evolve in time along with, or better ahead of fraud-detection mechanisms. This cat-and-mouse play between fraudsters and fraud fighters may seem to be an endless game, yet there is no alternative solution so far. By adopting and developing advanced fraud-detection and prevention mechanisms, organizations do manage to reduce losses due to fraud because fraudsters, like other criminals, tend to look for the easy way and will look for other, easier opportunities. Therefore, fighting fraud by building advanced and powerful detection systems is definitely not a pointless effort, but admittedly, it is very likely an effort without end.

Fraud is often as well a carefully organized crime, meaning that fraudsters often do not operate independently, have allies, and may induce copycats. Moreover, several fraud types such as money laundering and carousel fraud involve complex structures that are set up in order to commit fraud in an organized manner. This makes fraud not to be an isolated event, and as such in order to detect fraud the context (e.g., the social network of fraudsters) should be taken into account. Research shows that fraudulent companies indeed are more connected to other fraudulent companies than to nonfraudulent companies, as shown in a company tax-evasion case study by Van Vlasselaer et al. (2015). Social network analytics for fraud detection, as discussed in Chapter 5, appears to be a powerful tool for unmasking fraud by making clever use of contextual information describing the network or environment of an entity.

A final element in the description of fraud provided by Van Vlasselaer et al. indicates the many different types of forms in which fraud occurs. This both refers to the wide set of techniques and approaches used by fraudsters as well as to the many different settings in which fraud occurs or economic activities that are susceptible to fraud. Table 1.1 provides a nonexhaustive overview and description of a number of important fraud types—important being defined in terms of frequency of occurrence as well as the total monetary value involved.

Table 1.1 Nonexhaustive List of Fra...