Clustering of textual inputs in large eParticipation projects

Abstract

eParticipation projects often comprise deliberation among participants, where proposals are collected, discussed and rated to be processed in further stages of the eParticipation. In the case of large-scale audiences, this initial stage involves mass-online deliberation (MOD), which has to cope with a potentially very large number of proposals advanced by the participants. To enable clustering, MOD rely on human- (i.e. participant)-based appraisals of proposals given in the course of the participation project. Based on these appraisals, this contribution then proposes a clustering algorithm that makes use of the whole set of the above individual ratings. (Dis-)approval ratings are first weighted by the indication of clarity, that is, the higher the clarity rating assigned by a person, the higher the weight with which the (dis)approval rating will enter the clustering.