This thesis investigates fraud and the situation on Internet with e-commerce today, to point on some potential threats and needed countermeasures. The work reviews several state of the art e-fraud schemes, techniques used in the schemes and statistics on the extent of e-fraud. This part shows that e-frauds are today both sophisticated and widespread. Since real world frauds are deemed impossible to fully cover in order to predict potential new e-frauds, the thesis adopts a different approach. It suggests two abstraction models for fraud cases, a protocol model and a functional model. These are used to perform analysis on case studies on both telecom frauds and e-frauds. The analysis presents characteristics for both types of frauds. Using one of the abstraction models, the functional model, conceptually similar cases among telecom frauds as well as e-fraud cases are identified. The similar cases in each category are then compared, using the other abstraction model, the protocol model. The study shows that concepts from telecom frauds already exist in e-frauds. Several challenges and some possibilities in e-fraud prevention and detection are also extracted in the comparative study of the different categories. The major consequence of the challenges is e-frauds’ higher scalability compared to telecom frauds. Finally, this thesis covers several existing countermeasures in e-commerce along with specific countermeasures against auction fraud, phishing and spam. However, it is shown that these countermeasures do not address the challenges in e-fraud prevention and detection to a satisfactory extent. Therefore, this thesis proposes several high-level countermeasures in order to address the challenges.

Context: Today it is easy to do banking transaction digitally, both on a computer or by using a mobile phone. As the banking-services increases and gets implemented to multi-platforms it makes it easier for a fraudster to commit financial fraud. This thesis will focus on investigating log-files from a Mobile Money system that makes it possible to do banking transactions with a mobile phone.  Objectives: The objectives in this thesis is to evaluate if it is possible to combine two statistical methods, Benford's law together with statistical quantiles, to find a statistical way to find fraudsters within a Mobile Money system. Methods: Rules was extracted from a case study with focus on a Mobile Money system and limits was calculated by using quantiles. A fraud detector was implemented that use these rules together with limits and Benford's law in order to detect fraud.The fraud detector used the methods both independently and combined.The performance was then evaluated. Results: The results show that it is possible to use the Benford's law and statistical quantiles within the studied Mobile Money system. It is also shown that there is only a very small difference when the two methods are combined or not both in detection rate and accuracy precision. Conclusions: We conclude that by combining the chosen methods it is possible to get a medium-high true positive rates and very low false positive rates. The most effective method to find fraudsters is by only using quantiles. However, combining Benford's law with quantiles gives the second best result.

Billions of dollars of loss are caused every year by fraudulent credit card transactions. The design of efficient fraud detection algorithms is key for reducing these losses, and more and more algorithms rely on advanced machine learning techniques to assist fraud investigators. The design of fraud detection algorithms is however particularly challenging due to the non-stationary distribution of the data, the highly unbalanced classes distributions and the availability of few transactions labeled by fraud investigators. At the same time public data are scarcely available for confidentiality issues, leaving unanswered many questions about what is the best strategy. In this thesis we aim to provide some answers by focusing on crucial issues such as: i) why and how undersampling is useful in the presence of class imbalance (i.e. frauds are a small percentage of the transactions), ii) how to deal with unbalanced and evolving data streams (non-stationarity due to fraud evolution and change of spending behavior), iii) how to assess performances in a way which is relevant for detection and iv) how to use feedbacks provided by investigators on the fraud alerts generated. Finally, we design and assess a prototype of a Fraud Detection System able to meet real-world working conditions and that is able to integrate investigators’ feedback to generate accurate alerts. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

Whereas the problem of fraud is a global one, the rate and extent to which it is perpetrated in Nigeria, particularly in the public sector, is quite high and alarming. Literature reveals that different fraud prevention and detection mechanisms are being adopted to combat the menace of fraud; forensic accounting techniques appears to be the most effective and are currently used in most developed countries of the world. However, the extent to which forensic accounting techniques are being applied in fraud prevention and detection in Nigeria is not known. Also, the intention to use forensic accounting services in the public service has not been investigated. This study was therefore aimed at examining the application of forensic accounting techniques in fraud prevention and detection in Nigeria. Specific objectives were: (1) to investigate the mechanisms of fraud prevention and detection, and their levels of effectiveness in Nigeria, (2) to identify the major factors that hinder the application of forensic accounting techniques in fraud prevention and detection in Nigeria, (3) to examine practitioners' opinions and behavioural intention to use forensic accounting techniques in fraud prevention and investigation in Nigeria, (4) to explore the level of awareness of forensic accounting techniques in Nigeria and (5) to examine the readiness of universities in taking up forensic accounting courses. The study involved the collection of quantitative data. These data were collected from three sets of populations, viz. accounting students, accounting academics and accounting practitioners. The questionnaire served as the survey instruments. The data collected were analysed using appropriate statistical techniques and computer software. The study identified several fraud prevention and detection mechanisms that are currently used in Nigeria, such as systems of internal controls, operational audits and corporate code of conduct. Students' t-test indicates a significant difference between the perceived effectiveness and actual usage of fraud prevention and detection mechanisms in Nigeria. It was further discovered that the most effective mechanisms, like the forensic accounting techniques, are the least used in fraud prevention and detection. This implies that the current mechanisms of fraud prevention and detection are not proactive in dealing with the fraud menace. Also, legal, educational and political factors were identified to hinder the application of forensic accounting techniques in Nigeria. The level of awareness in forensic accounting in Nigeria is generally low. While the one-way analysis of variance indicates a significant variation among the three populations, it was discovered that students had the lowest level of awareness. Further findings of the study reveal that the universities are not yet ready to take up forensic accounting courses. Using the structural equation modelling (SEM), all the other seven propositions were supported. The findings of this study have both theoretical and practical implications. Theoretically, it further strengthened the findings of previous studies on the organisational intention. From the practical point of view, there is urgent need for manpower development in universities with specialisation in forensic accounting. Again, the educational institutions, and particularly the universities in Nigeria, would need to include forensic accounting courses in the undergraduate curriculum as education has been shown to be pivotal in creating awareness on the use of forensic accounting techniques. Furthermore, from the sampled space, the study has captured the current state of forensic accounting in Nigeria and the findings will be very useful for the public service, private organisations and policy makers.

The perpetration of accounting fraud still remains a prevalent and significantly costly issue in today's business world. The names Enron, WorldCom, HealthSouth, and Madoff are still all too recent reminders of the devastating cost of financial statement fraud. Management accountants, as preparers of these statements, are in the best position to detect such fraud. Yet there exists no current measurement instrument or methodology designed to measure a management accountant's intention to report fraud. The primary purpose of this study was to investigate the beliefs, concepts, and antecedents that provide the motivation to, or the deterrent from, the reporting of fraudulent accounting activity when witnessed by professional management accountants, and develop an instrument that might measure that motivation. The theoretical basis that framed this research was the Theory of Planned Behavior which provides for an analysis of a participant's attitude, subjective norm, and perceived behavioral control in the development of the intention to perform a specific behavior. The population studied was the U.S. membership of the Institute of Management Accountants, and grant assistance and support was provided by the Institute's Research Foundation. The sample from this population formed a very appropriate representation of experienced, professional management accountants. . No previous research involving this population with the application of the Theory of Planned Behavior and the investigation of the reporting of fraudulent accounting activity had been conducted. Therefore, there were no existing survey instruments that could be applied. The development of an original survey questionnaire to specifically address this research was required. The distribution of this survey questionnaire resulted in 285 complete and usable responses. These responses measured the strength of the participant's positive or negative beliefs concerning the antecedents related to the three exogenous constructs of the Theory of Planned Behavior - attitude, subjective norm, and perceived behavioral control, and the endogenous construct of intention. Structural Equation Modeling (SEM) with measured variables was chosen as the methodology for the analysis of the results measured in the survey responses. Confirmatory Factor Analysis was applied to each construct individually, and construct items were modified to obtain the most reasonable model fit, validity, and reliability. Items were combined into composites to represent the constructs of interest in the theory, as measured by the survey. The relations among the constructs of the Theory of Planned Behavior were then specified using these composites in an SEM model. The results of the data and the findings of the SEM model indicated that professional management accountants form a strong positive intention to report the witnessing of accounting fraud. The positive beliefs that formed the exogenous variables that showed statistically significant effects on the endogenous variable of the formation of a positive intention to report fraudulent accounting activity were: support of the system of internal control, prevention of financial loss, retention of the integrity and ethical values of the profession, perceived support of significant others, and limited impediment due to fear of retaliation. A surprising result was that 32% of all respondents indicated a lack of easy/any access to an anonymous fraud reporting hotline, which is an issue for further research. This study provides additional insight into the concepts, beliefs, and antecedents that form a professional management accountant's intention to report fraudulent accounting activity. The study also presents the basis of a preliminary instrument for the measurement of the intention of management accountants to report fraudulent accounting activity. Further research is suggested for the identification of additional concepts, antecedents, and beliefs related to fraud reporting and for the development of an even more effective measurement instrument.

In “web-based scams”, scam websites provide fraudulent business or fake services to steal money and sensitive information from unsuspecting victims. Despite the researchers’ efforts to develop anti-scam detection techniques, the scams continue to evolve and cause online threats. State-of-the-art anti-scam research still faces several challenges, such as automatically acquiring a labeled scam dataset and providing early detection and prevention mechanisms to attacks that use cryptocurrency as a payment medium. In this thesis, we implement a data-driven model to detect and track web-based scams with a web presence. Given a few scam samples, our model formulates scam-related search queries and uses them on multiple search engines to collect data about the websites to which victims are directed when they search online for sites that may be related to the scam. After collecting a sufficient corpus of web pages, our model semi-automatically clusters the search results and creates a labeled training dataset with minimal human interaction. Our model proactively looks for scam pages and monitors their evolution over time rather than waiting for the scam to be reported. Whenever a new scam instance is detected, the model sends it automatically to the eCrime eXchange data warehouse in real-time. We have used the model to investigate and gain knowledge on two scams; the “Game Hack” Scam (GHS) and the “Bitcoin Generator Scam” (BGS). To the best of our knowledge, GHS and BGS have not been well studied so far, and this is the first systematic study of both scams. GHS targets game players, in which the attackers attempt to convince victims that they will be provided with free in-game advantages for their favorite game. Before claiming these advantages, the victims are supposed to complete one or more tasks, such as filling out “market research” forms and installing suspicious executable files on their machines. Over a year of crawling, we uncovered more than 5,900 unique domains. We estimate that these domains have been accessed at least 150 million times from 2014 until 2019. BGS is a simple system in which the scammers promise to “generate” new bitcoins using the ones sent to them. BGS is not a very sophisticated attack; the modus operandi is to put up some web page that contains the address to send the money and wait for the payback. Over 21 months of crawling, we found more than 3,000 addresses directly associated with the scam, hosted on over 1,200 domains. Overall, these addresses have received (at least) over 9.6 million USD. Our analysis showed that a small group of scammers controls the majority of the received funds. The top two groups have received around 6 million USD, which is more than half of the total funds received by the scam addresses.

Microfinance institutions (MFIs) are an important tool of poverty reduction which has gained grounds over the years and grown rapidly given the services they provide. The rapid growth of the MFIs has had huge challenges on their regulatory framework which in turn has resulted in the prevalence of fraudulent cases. With the devasting effects of fraud on MFIs and the importance of MFIs in many economies this thesis aims to examine how the design and use of internal control impact fraud detection and prevention in MFIs. To achieve this aim, a qualitative study was conducted with a case study on eight MFIs in Cameroon. Primary data will be obtained from fourteen semi-structured interviews. Data will be analysed manually using thematic analysis. The findings revealed that internal control has a positive impact on fraud detection and prevention in MFIs by reducing fraud incentive, opportunity, rationalization, and capability. Further, findings revealed that the greatest causes of fraud in MFIs are poor remuneration, weak monitoring, and a poor internal control system. Therefore, for the purpose of future fraud prevention, MFIs should ensure to improve their remuneration schemes, improve1 their monitoring system and ensure regular internal control system updates in term of software and design. The study also, suggests further research on this topic in MFIs with a case study in other countries of the world. It will also be interesting for other researchers to explore how the aspect of capability as a key determinant of fraud can be reduced. This thesis contributes to academic literature as there is lack of studies on the impact of internal control on fraud detection and prevention in MFIs.

Fraud is a threat that most online service providers must address in the development of their systems to ensure an efficient security policy and the integrity of their revenue. Amadeus, a Global Distribution System providing a transaction platform for flight booking by travel agents, is targeted by fraud attempts that could lead to revenue losses and indemnifications. The objective of this thesis is to detect fraud attempts by applying machine learning algorithms to bookings represented by Passenger Name Record history. Due to the lack of labelled data, the current study presents a benchmark of unsupervised algorithms and aggregation methods. It also describes anomaly detection techniques which can be applied to self-organizing maps and hierarchical clustering. Considering the important amount of transactions per second processed by Amadeus back-ends, we eventually highlight potential bottlenecks and alternatives.

The lack of legitimate datasets on mobile money transactions toperform research on in the domain of fraud detection is a big prob-lem today in the scientic community. Part of the problem is theintrinsic private nature of mobile transactions, not much infor-mation can be exploited. This will leave the researchers with theburden of rst harnessing the dataset before performing the actualresearch on it. The dataset corresponds to the set of data in whichthe research is to be performed on. This thesis discusses a solutionto such a problem, namely the Paysim simulator. Paysim is a -nancial simulator that simulates mobile money transactions basedon an original dataset. We present a solution to ultimately yieldthe possibility to simulate mobile money transactions in such a waythat they become similar to the original dataset. The similarity orthe congruity will be measured by calculating the error-rate betweenthe synthetic data set and the original data set. With technologyframeworks such as "Agent Based" simulation techniques, and theapplication of mathematical statistics, it can be demonstrated thatthe synthetic data is as prudent as the original data set. The aimof this thesis is to demonstrate with statistical models that PaySimcan be used as a tool for the intents of nancial simulations.

The aim of explainable machine learning is to aid humans in understanding how exactly complex machine learning models work. Machine learning models have offered great performance in various areas. However, the mechanisms behind how the model works and how decisions are being made remain unknown. This specific constraint increases the user’s hesitation to trust the results of the model and even to improve their performance further. Counterfactual explanation is one method to offer explainability in machine learning by indicating what would have happened if the input of a model was modified in a specific way. Fraud is the action of acquiring something from someone else in a dishonest manner. Companies’ and organizations’ vulnerability to malicious actions has been increasing due to the development of digitalization. Machine learning applications have been successfully put in place to tackle fraudulent actions. However, the severity of the impact of fraudulent actions has highlighted the need for further scientific exploration of the topic. The current research will attempt to do so by studying counterfactual explanations related to fraud detection. Latent-CF is a method for counterfactual generation that utilizes an autoencoder and gradient descent in its latent space. LatentCF++ is an extension of Latent-CF. It utilizes a classifier and an autoencoder. The aim is to perturb the encoded latent representation through a gradient descent optimization for counterfactual generation so that the initially undesired class is then classified with the desired prediction. Compared to Latent-CF, LatentCF++ uses Adam optimization and adds further constraints to ensure that the generated counterfactual’s class probability surpasses the set decision boundary. The research question the current thesis addresses is: “To what extent can LatentCF++ provide reliable counterfactual explanations in fraud detection?”. In order to provide an answer to this question, the study is applying an experiment to implement a new application of LatentCF++. The current experiment utilizes a onedimensional convolutional neural network as a classifier and a deep autoencoder for counterfactual generation in fraud data. This study reports satisfying results regarding counterfactual explanation production of LatentCF++ on fraud detection. The classification is quite accurate, while the reconstruction loss of the deep autoencoder employed is very low. The validity of the counterfactual examples produced is lower than the original study while the proximity is lower. Compared to baseline models, k-nearest neighbors outperform LatentCF++ in terms of validity and Feature Gradient Descent in terms of proximity.