Annex: use of the monitoring tools in WebAriadne

AuthorFrancesca Torti - Marco Riani - Anthony C. Atkinson - Domenico Perrotta - Aldo Corbellini
ProfessionEuropean Commission, Joint Research Centre (JRC) - University of Parma, Italy - London School of Economics, UK - European Commission, Joint Research Centre (JRC) - University of Parma, Italy
Pages31-33
B. Annex: use of the monitoring tools in WebAriadne
Our SAS library for robust regression is motivated by problems linked to the Customs Union and Anti-Fraud
policies of the European Union (EU), which are rooted in its founding Treaties TFEU (2012). In fact, the
ef‌f‌icient implementation of such policies, as also stated by the World Customs Organization WCO (2017),
calls for a modernization of the national anti-fraud services through the adoption of tools based on state-
of-the-art mathematical, statistical and computer science methods. In this framework, the Joint Research
Centre (JRC) of the European Commission delivers such tools to the law-enforcement partners in the EU
Institutions and Member States since decades (the roots of the activity can be dated back to 1995). The role
assigned to the JRC in this policy domain comprises the modeling of fraud in pertinent statistical data, the
development of the related statistical methods for fraud detection, their product software implementation,
their deployment as services accessible to customers, and the routine dissemination of alerts (fraud relevant
signals) to authorized users through the web. More precisely, the users access alerts related to trade-based
illicit activities through the THESEUS resource (https://theseus.jrc.ec.europa.eu) or generate them
in full autonomy, on data of their choice, using tools accessible through the web application WebARIADNE
(https://webariadne.jrc.ec.europa.eu). Figure 17 shows their login page. The SAS library discussed
in the paper is a key module of both THESEUS and WebARIADNE, illustrated below with few snapshots.
The left panel of Figure 18 shows how the user selects a dataset of interest and the statistical application
to apply. On the right panel the user has selected a local dataset and is presented with a preview of its
content. The user might also want to analyze a previously uploaded dataset: Figure 19 shows the preview
given when the dataset is selected, with the list of the f‌ields and a sample of records. The frame of
Figure 20 is specif‌ic to the SAS-based regression module of this paper. Here, the user can set the key
input parameters, namely the estimation method for f‌itting the data (FS, LTS, LMS, S, M, MM), the
signif‌icance level for detecting the outliers, the dependent and independent regression variables, possible
grouping variables for partitioning the dataset in homogeneous groups. Figure 21 shows the output of the
most recent runs: there are three runs with the Forward Search and one with Least Trimmed Squares. Such
output is typically presented to the user as a table with the relevant input variables and generated statistics
(estimated parameters, pvalues, residuals, etc.). Same results are also disseminated to the authorized users
of THESEUS in a comprehensive form: an example is shown in Figure 22.
Figure 17: The login page of the WebARIADNE and THESEUS applications.
31

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT