Data Mining for Intelligence, Fraud & Criminal Detection

Description

In 2004, the Government Accountability Office provided a report detailing approximately 200 government-based data-mining projects. While there is comfort in knowing that there are many effective systems, that comfort isn’t worth much unless we can determine that these systems are being effectively and responsibly employed. Written by one of the most respected consultants in the area of data mining and security, Data Mining for Intelligence, Fraud & Criminal Detection: Advanced Analytics & Information Sharing Technologies reviews the tangible results produced by these systems and evaluates their effectiveness. While CSI-type shows may depict information sharing and analysis that are accomplished with the push of a button, this sort of proficiency is more fiction than reality. Going beyond a discussion of the various technologies, the author outlines the issues of information sharing and the effective interpretation of results, which are critical to any integrated homeland security effort. Organized into three main sections, the book fully examines and outlines the future of this field with an insider’s perspective and a visionary’s insight. Section 1 provides a fundamental understanding of the types of data that can be used in current systems. It covers approaches to analyzing data and clearly delineates how to connect the dots among different data elements Section 2 provides real-world examples derived from actual operational systems to show how data is used, manipulated, and interpreted in domains involving human smuggling, money laundering, narcotics trafficking, and corporate fraud Section 3 provides an overview of the many information-sharing systems, organizations, and task forces as well as data interchange formats. It also discusses optimal information-sharing and analytical architectures Currently, there is very little published literature that truly defines real-world systems. Although politics and other factors all play into how much one agency is willing to support the sharing of its resources, many now embrace the wisdom of that path. This book will provide those individuals with an understanding of what approaches are currently available and how they can be most effectively employed. Review: An excellent source of insight into the analysis of fraud, money laundering and financial crime - Fraud and related crime can be a dry topic. But equally, money underpins the significant majority of other crimes. It follows that most criminal investigators and analysts ought to have a sound knowledge and understanding of how fraud and related financial crimes work. The author has done an excellent job at making this topic very approachable for both those new to it and those not so new. The author includes substantial coverage of AMLCTF issues and addresses local, national and transnational criminal activity. Areas covered include warranty fraud, Medicare fraud, expense fraud, cheque kiting, bait & switch, tax fraud, identity fraud, drug trafficking finance and terrorism. Indeed, there is something in this book to interest almost every criminal investigator. That said, no book can be all things to all men, and some areas inevitably have to miss out. I would personally have liked to see some coverage of divorce fraud, as it is a growth area for lawyers, but it is still niche so perhaps in the next edition. I would also have liked a bit more variety in the illustrations that the author used. Perhaps the use of different analytical software packages to break up the monotony. I would also have liked to have seen some coverage of computer forensics, data mining and 'big data' visualization, as these are emerging fields and arguably define the future of "advanced analytics". But, again, perhaps these will receive coverage in the next edition. And there needs to be a next edition, because this is an area of crime that is continually changing. But these are very minor issues, especially when compared to the overall quality of the content of the book. I envisage this book becoming a definitive text, much like many of the well known legal texts that have gone on to have multiple editions published over many decades. This book is worth every one of its five stars. Review: Great Advice and Information, but No Advanced Analytics - Christopher Westphal has created a very engaging and informative book about various ways that criminals lie, cheat, and steal, and he has presented some intriguing suggestions for how to catch them. While many of his sidebars about topics like social security numbers, ICD9 codes, and vehicle identification numbers (VINs) are pulled almost word-for-word from the Web, it is convenient to have the information in one easy-to-read source, and the links are cited if you want to dig more deeply into a topic. I highly recommend the book for anyone interested in learning about fraud detection. However, I only give the book four stars out of five because despite the subtitle, there are absolutely no advanced analytics in the book. In fact, there are no analytics in the book beyond querying a database. There is a common refrain that goes, "Data mining is NOT querying and reporting." Thus, I would even question the use of the term "data mining" in the title. Data mining does include visualization of data, and this is where the book excels. The visual displays of data certainly enhance the learning experience. However, the visuals usually just represent summary statistics extracted from a relational database. If you look for evidence of "advanced analytics" in the index, you will struggle. None of the following common advanced analytic techniques are mentioned in the index, and I could find no mention of them in the text as well: decision trees (recursive partitioning), neural networks, generalized additive models, case-based reasoning, regression models, MARS, LARS, support vector machines. Some data mining references will pick one technology, like decision trees, and focus on that one technique. I could not identify a single advanced analytic technique anywhere in the book, and certainly not in the index. Even the generic terms "supervised learning" and "unsupervised learning" are nowhere to be found. I looked at the index online at desertcart before I bought the book, so I am not surprised nor disappointed with what I received. In fact, I had relatively low expectations for the book, but found Westphal more than delivered in describing details of warranty fraud, money laundering, gift card fraud, and many other fraud topics. I am an expert on insurance fraud detection, and I found Westphal to be well informed in that area. Furthermore, I found the book to be ENTERTAINING, a rarity in a book with "data mining" in the title. I have found many analytic experts who struggle in applying analytics because they do not understand the business problem. I would recommend the book for anyone with analytical skills who is new to any of the areas that Westphal addresses. In summary, if the book had a more accurate title, I would probably give it five stars.