Detection of Anomalous Activity in Mobile Money Transfer Services Using RadViz-Visualization

Nowadays, mobile communication networks represent a key enabling infrastructure for financial service provision, since they offer significant opportunities for increasing the efficiency and pervasiveness of such services by expanding access and lowering transaction costs. In the paper, the authors analyze the use case of mobile money transfer services which are managed by a mobile network operator who not only provides infrastructure to financial services but also emits mobile money.
In this paper, we present an interactive multi-view visualization approach that provides a better insight in the large data sets describing MMTS activity. It is based on a RadViz-related visualization of the MMTS users that helps to determine groups of similarities and outliers among them and is characterized by low computational complexity. To the best of knowledge of the authors, this work is the first to exploit the RadViz-visualization technique to visualize MMTS subscribers. RadViz –based presentation of the MMTS users is supported by interactive graph based visualization of their contacts. The graph of the users’ contacts is often used to analyze financial transactions as it allows discovering structural peculiarities such as bridges and cliques.
The proposed visual analytics technique was evaluated on different test data sets containing different fraudulent financial scenarios. Summarizing the results of the efficiency evaluation of the proposed visualization technique for MMTS transaction activity, we can say that RadViz visualization is helpful when detecting fraudulent scenarios which make use of mules users whose behavior significantly differs from the behavior of the other MMTS subscribers. It also allows detecting frauds associated with shifts in user behavior which have cumulative character. Thus these frauds can be revealed when choosing a relatively long period of time (e.g. a month) to explore MMTS transactions. That is why this technique was effective when detecting mules in the money laundering scheme and the mobile botnet.