Analysing Political Opinions Using Redescription Mining

Esther Galbrun (1) , Pauli Miettinen (2)

1 ORPAILLEUR - Knowledge representation, reasonning
2 Max Planck Institute for Informatics [Saarbrücken]

Understanding the socio-economical background of voters supporting a certain cause or, vice versa, understanding the political stance of people from a certain socio-economical niche are important questions in political sciences. Traditionally, answering these questions has required the researcher to fix either the political stance or the socio-economical background. In this paper, we propose using redescription mining to automatically find the stances and niches that correspond to each other. We show how redescription mining can be applied to open data from voting advice applications, providing insights about the position of the candidates to parliamentary elections. Furthermore, we show that these insights are not only descriptive, but that they also generalize well to new data.

Domaines: Intelligence artificielle [cs.AI]

Esther Galbrun, Pauli Miettinen. Analysing Political Opinions Using Redescription Mining. Proceedings of the Data Mining in Politics worshop at ICDM 2016, DMiP'16, Dec 2016, Barcelona, Spain. ⟨hal-01399254⟩