Entropy based probabilistic collaborative clustering

Jérémie Sublime, Basarab Matei, Guénaël Cabanes, Nistor Grozavu, Younès Bennani, Antoine Cornuéjols

Abstract: Highlights•A new collaborative framework is proposed for unsupervised learning.•This framework allows algorithms of different families to work together.•The collaboration process relies on the solutions matrices and their diversity.•The convergence of the method is proved using variational EM.