Véronique Van Vlasselaer

APATE: A novel approach for automated credit card transaction fraud detection using network-based extensions

In the last decade, the ease of online payment has opened up many new opportunities for e-commerce, lowering the geographical boundaries for retail. While e-commerce is still gaining popularity, it is also the playground of fraudsters who try to misuse the transparency of online purchases and the transfer of credit card records. This paper proposes APATE, a novel approach to detect fraudulent credit card transactions conducted in online stores. Our approach combines (1) intrinsic features derived from the characteristics of incoming transactions and the customer spending history using the fundamentals of RFM (Recency - Frequency - Monetary); and (2) network-based features by exploiting the network of credit card holders and merchants and deriving a time-dependent suspiciousness score for each network object. Our results show that both intrinsic and network-based features are two strongly intertwined sides of the same picture. The combination of these two types of features leads to the best performing models which reach AUC-scores higher than 0.98.

GOTCHA! Network-Based Fraud Detection for Social Security Fraud

We study the impact of network information for social security fraud detection. In a social security system, companies have to pay taxes to the government. This study aims to identify those companies that intentionally go bankrupt to avoid contributing their taxes. We link companies to each other through their shared resources, because some resources are the instigators of fraud. We introduce GOTCHA!, a new approach to define and extract features from a time-weighted network and to exploit and integrate network-based and intrinsic features in fraud detection. The GOTCHA! propagation algorithm diffuses fraud through the network, labeling the unknown and anticipating future fraud while simultaneously decaying the importance of past fraud. We find that domain-driven network variables have a significant impact on detecting past and future frauds and improve the baseline by detecting up to 55% additional fraudsters over time. This paper was accepted by Lorin Hitt, information systems.

Guilt-by-Constellation: Fraud Detection by Suspicious Clique Memberships

Given a labeled graph containing fraudulent and legitimate nodes, which nodes group together? How can we use the riskiness of node groups to infer a future label for new members of a group? This paper focuses on social security fraud where companies are linked to the resources they use and share. The primary goal in social security fraud is to detect companies that intentionally fail to pay their contributions to the government. We aim to detect fraudulent companies by (1) propagating a time-dependent exposure score for each node based on its relationships to known fraud in the network, (2) deriving cliques of companies and resources, and labeling these cliques in terms of their fraud and bankruptcy involvement, and (3) characterizing each company using a combination of intrinsic and relational features and its membership in suspicious cliques. We show that clique-based features boost the performance of traditional relational models.

Using social network knowledge for detecting spider constructions in social security fraud

As social networks offer a vast amount of additional information to enrich standard learning algorithms, the most challenging part is extracting relevant information from networked data. Fraudulent behavior is imperceptibly concealed both in local and relational data, making it even harder to define useful input for prediction models. Starting from expert knowledge, this paper succeeds to efficiently incorporate social network effects to detect fraud for the Belgian governmental social security institution, and to improve the performance of traditional non-relational fraud prediction tasks. As there are many types of social security fraud, this paper concentrates on payment fraud, predicting which companies intentionally disobey their payment duties to the government. We introduce a new fraudulent structure, the so-called spider constructions, which can easily be translated in terms of social networks and included in the learning algorithms. Focusing on the egonet of each company, the proposed method can handle large scale networks. In order to face the skewed class distribution, the SMOTE approach is applied to rebalance the data. The models were trained on different timestamps and evaluated on varying time windows. Using techniques as Random Forest, logistic regression and Naive Bayes, this paper shows that the combined relational model improves the AUC score and the precision of the predictions in comparison to the base scenario where only local variables are used.

AFRAID: Fraud Detection via Active Inference in Time-evolving Social Networks

Fraud is a social process that occurs over time. We introduce a new approach, called AFRAID, which utilizes active inference to better detect fraud in time-varying social networks. That is, classify nodes as fraudulent vs. non-fraudulent. In active inference on social networks, a set of unlabeled nodes is given to an oracle (in our case one or more fraud inspectors) to label. These labels are used to seed the inference process on previously trained classifier(s). The challenge in active inference is to select a small set of unlabeled nodes that would lead to the highest classification performance. Since fraud is highly adaptive and dynamic, selecting such nodes is even more challenging than in other settings. We apply our approach to a real-life fraud data set obtained from the Belgian Social Security Institution to detect social security fraud. In this setting, fraud is defined as the intentional failing of companies to pay tax contributions to the government. Thus, the social network is composed of companies and the links between companies indicate shared resources. Our approach, AFRAID, outperforms the approaches that do not utilize active inference by up to 15% in terms of precision.

A Models Comparison to Estimate Commuting Trips Based on Mobile Phone Data

Upon an overall human mobility behavior within the city of Rio de Janeiro, this paper describes a methodology to predict commuting trips based on the mobile phone data. This study is based on the mobile phone data provided by one of the largest mobile carriers in Brazil. Mobile phone data comprises a reasonable variety of information about subscribers’ usage, including time and location of call activities throughout urban areas. This information was used to build subscribers’ trajectories, describing then the most relevant characteristics of commuting over time. An Origin-Destination (O-D) matrix was built to support the estimation for the number of commuting trips. Traditional approaches inherited from transportation systems, such as gravity and radiation models – commonly employed to predict the number of trips between locations(regularly upon large geographic scales) – are compared to statistical and data mining techniques such as linear regression, decision tree and artificial neural network. A comparison of these models shows that data mining models may perform slightly better than the traditional approaches from transportation systems when historical information are available. In addition to that, data mining models may be more stable for great variances in terms of the number of trips between locations and upon different geographic scales. Gravity and radiation models work very well based on large geographic scales and they hold a great advantage, they are much easier to be implemented. On the other hand, data mining models offer more flexibility in incorporating additional attributes about locations – such as number of job positions, available entertainments, schools and universities posts, among others –and historical information about the trips over time.