2019-04 Graduation Assignments Cyber Security

Measuring and analysing fraud victimization with a focus on online fraud.

Topic. There are not many adequate, systematic methods of measuring or understanding fraud in the world, including the Netherlands. To combat fraud, understanding the extent of the problem, how fraud works, and issues that lead to victimization is vital. Therefore, an adequate description of how fraud works, and its extent, is urgently required. We propose to develop a questionnaire to measure fraud victimization, based on literature, and collect data on fraud victimization. Existing questionnaires and Research on fraud victimization indicates that context matters for results (Beals, 2017). Additionally, research shows that careful questioning produces better data (Smith & Hutchings, 2014). This research would focus on designing and validating a new questionnaire focused on online fraud, with an eye on additional questions compared to existing measuring instruments, such as questions about monetary losses, reporting to the police or others or possible ideas about the offender to give us better data for further analysis and understanding of fraud.

Required expertise. We are looking for a master student with an interest in deviant behaviour and measurement issues and who is not afraid of smart data-collection and analysis.

Contact UT. Are you interested in writing your thesis on this topic? Please mail us.

·         Prof. Dr. Marianne Junger - IEBIS, BMS, Email: M.Junger@Utwente.nl

·         Dr. Guido van Capelleveen  - IEBIS, BMS, Email: g.c.vancapelleveen@utwente.nl   

·         Roeland Kegel – SCS, EWI, Email: r.h.p.kegel@utwente.nl 

References

Beals, M. E., Carr, D. C., Mottola, G. R., Deevy, M. J., & Carstensen, L. L. (2017). How Does Survey Context Impact Self-reported Fraud Victimization? The Gerontologist, 57(2), 329-340.

Smith, R. G., & Hutchings, A. (2014). Identity crime and misuse in Australia: Results of the 2013 online survey. Canberra, Australia: Australian Government. Australian Institute of Criminology. http://aic.gov.au/media_library/publications/rpp/128/rpp128.pdf.

Preventing fraud and developing prevention models

Topic.  The aim of this study is to develop models for the prevention of fraud according to the principles of Situational Crime Prevention (SCP) (Clarke, 1980). For effective prevention models, it is necessary to describe the modus operandi (Crime Scripts) of specific fraud types accurately. It can be also be helpful to describe the differences between resilient and susceptible people with regards to fraud types, as this can help identify measures that can be taken to assist vulnerable populations in recognizing fraud and avoiding victimization. Because the modus operandi of crimes are specific to a particular offense, SCP is always crime-specific. Using data provided by the FraudeHelpdesk (https://www.fraudehelpdesk.nl/), we want to identify the differences between online and offline fraud.

Required expertise. We are looking for a master student with an interest in cybersecurity and an interest in data analysis.

Contact UT. Are you interested in writing your thesis on this topic? Please mail us.

·         Prof. Dr. Marianne Junger - IEBIS, BMS, Email: M.Junger@Utwente.nl

·         Dr. Guido van Capelleveen  - IEBIS, BMS, Email: g.c.vancapelleveen@utwente.nl   

·         Roeland Kegel – SCS, EWI, Email: r.h.p.kegel@utwente.nl   

References

Clarke, R. V. (1980). Situational crime prevention: Theory and practice. Brit. J. Criminology, 20.

Experiment: preventing fraud based on financial intelligence

Topic. It is difficult to catch fraudsters, especially those who work online. Accordingly, to prevent crime it is often smarter to focus on (potential) victims. We propose, in collaboration with banks and possibly other financial agencies, to replicate an Australian project: ‘Operation/Project Sunbird[1]. In this operation, the Western Australian Police Major Fraud Squad and a Consumer Protection agency looked at who was sending money from Western Australia to African countries such as Benin, Ghana, Nigeria, Sierra Leone, and Togo. After an analysis of these transactions, potential fraud victims are identified, and they receive a letter warning them that they may become a victim of fraud. If no reaction of the victim is registered, they receive a second warning letter. This project was quite successful and most victims stopped sending money.

Between March 2013 and August 2015, approximately $51.5 million ($51,471,093.79) was sent from West Australia to West Africa. Of those funds, approximately $17.9 million ($17,857,507.43) was identified as the result of fraudulent transactions. Between March 2013 and October 2015, 3001 first letters and 1503 second letters were sent out. Monitoring of recipients occurs three months prior to sending a letter, excludes the month that the letter is sent, and then continues for three months after. Based on this, of those senders that received the first letter between March 2013 and July 2015, 73% stopped sending money and 13% reduced the amount of money sent. Of those who received a second letter during this same time period, 53% stopped sending money and 27% reduced the amount of money sent. A small percentage of individuals also continued to send money despite the intervention.”(Cross, 2016, p. 133).

Required expertise. We are looking for a master student with an interest in cultivating cooperation with financial institutions and an interest in big data analysis.

Please note: For this project we need the cooperation of the banks.

Contact UT. Are you interested in writing your thesis on this topic? Please mail us.

·         Prof. Dr. Marianne Junger - IEBIS, BMS, Email: M.Junger@Utwente.nl

·         Dr. Guido van Capelleveen  - IEBIS, BMS, Email: g.c.vancapelleveen@utwente.nl   

·         Roeland Kegel – SCS, EWI, Email: r.h.p.kegel@utwente.nl 

Exploring the “Dark Number” of fraud

FOR THIS PROJECT WE NEED THE COOPERATION OF CBS OR A SIMILAR INSTITUTE FOR DATA COLLECTION

Topic. Many people fall victim to some form of fraud every day. But not all cases of fraud are reported. The causes for this range from a fear of admitting to having been ‘duped’, to a lack of faith in the chances that the perpetrator can be caught or that any of the stolen resources (money, identity, trust) can be recovered. The lack of accurate data hampers any insights into assessing the societal relevance, damages and trends in fraud.

Assessing the magnitude of the number of unreported fraud cases can help in fraud prevention and analysis. To do this, we might take several types of fraud and develop a questionnaire to be put to a sample of participants, examining the frequency of and reasons behind non-disclosure of fraud victimization, specifically.

Required expertise. We are looking for a master student with an interest in understanding victim behaviour and is motivated to look for new ways to approach a tricky data analysis problem.

Contact UT. Are you interested in writing your thesis on this topic? Please mail us.

·         Prof. Dr. Marianne Junger - IEBIS, BMS, Email: M.Junger@Utwente.nl

·         Roeland Kegel – SCS, EWI, Email: r.h.p.kegel@utwente.nl 

·         Dr. Guido van Capelleveen  - IEBIS, BMS, Email: g.c.vancapelleveen@utwente.nl  

Examining Fraud Mentions in Social Media

Topic. Fraud, especially digital fraud, is a booming criminal enterprise. Every year, new types of scams are developed and used to defraud victims for significant amounts of money. In fact, since crime scripts evolve so rapidly, it has become hard to keep up with the latest trends and techniques that criminals use.

A possible way to understand what is going on right now might be estimated by examining mentions of fraud in social media. By mining information from hashtags on twitter, we might be able to discern existing, or even new types of fraud in the process of emerging.

Required Expertise. We are looking for a computer science or a BIT master student who likes programming and data analysis, with an interest in understanding the latest in the social aspects of Cybercrime.

Contact UT. Are you interested in writing your thesis on this topic? Please mail us.

·         Roeland Kegel – SCS, EWI, Email: r.h.p.kegel@utwente.nl 

·         Prof. Dr. Marianne Junger - IEBIS, BMS, Email: M.Junger@Utwente.nl

·         Dr. Guido van Capelleveen  - IEBIS, BMS, Email: g.c.vancapelleveen@utwente.nl  

Detecting “pop-up web-shops”

Topic. A popular kind of online market fraud involves the creation of a new web-shop where customers can buy items. After customers place their orders and pay, the web-shop disappears without delivering any of the promised goods. These web-shops typically only stay online for a few days (hence the name “pop-up web-shop”).

We might understand how recognize such a web-shop by analysing the content of several such web-shops, which can help us in developing tools that can combat this new digital threat. One possible approach to finding such web-shops would be to identify newly registered domains and crawling them to find suspicious examples.

Required expertise. We are looking for a computer science or a BIT master student who is interested in the technical aspects of detecting fraud in progress, that is not afraid of either programming or machine learning.

Contact UT. Are you interested in writing your thesis on this topic? Please mail us.

·         Roeland Kegel – SCS, EWI, Email: r.h.p.kegel@utwente.nl 

·         Prof. Dr. Marianne Junger - IEBIS, BMS, Email: M.Junger@Utwente.nl

·         Dr. Guido van Capelleveen  - IEBIS, BMS, Email: g.c.vancapelleveen@utwente.nl  

‘fingerprinting’ fraud

Topic

It is hard to catch fraudsters. Single fraud cases are too time consuming to investigate exhaustively, but it may be worthwhile to investigate a class of many similar fraud attempts. So our question in this project is: is it possible to make ‘fingerprints’ of fraud, based on reports from the FraudeHelpdesk (https://www.fraudehelpdesk.nl/)?

If so, can one automatically generate recommendations based on individual fraud reports that show similar patterns to other reports, in order to recognize a group of incidents that can lead us to a single (or group of) fraudster(s)? Or even link different types of fraud to a group of fraudsters. If you are interested in creating a fraud classification system based on fraud report data, this is a project for you.

Required expertise: We are looking for a computer science or a BIT master student who is interested in the technical aspects of detecting fraud in progress that has an interest both in programming and machine learning techniques.

Contact UT. Are you interested in writing your thesis on this topic? Please mail us.

·         Dr Guido van Capelleveen  - IEBIS, BMS, Email: g.c.vancapelleveen@utwente.nl  

·         Roeland Kegel – SCS, EWI, Email: r.h.p.kegel@utwente.nl 

·         Prof Dr Marianne Junger - IEBIS, BMS, Email: M.Junger@Utwente.nl

[1] https://www.scamnet.wa.gov.au/scamnet/About_us-Media_and_events-OperationProject_Sunbird.htm