Analyse de comportements relatifs exceptionnels expliquée par des textes : les votes du parlement européen.
Abstract
We use DEBuNk , a data mining algorithm which enables identifying groups and contexts
showing a relative behavior that is exceptional, compared to the usual relative behavior of these
groups. For a better understanding and interpretation of the results, the visual interface includes
a text generation module which transcribes the obtained patterns into texts. We illustrate our
approach through the web platform ANCORE with voting data from the European Parliament.