I am interested in machine learning, graph theory, lattices and complexity, my thesis deals with learning, modeling, and preferences prediction. I focus on conditional preferences networks, also called CP-nets, and their learning when data are corrupted by noise. I also work on formal concept analysis (FCA) in order to find a link between lattices and CP-nets (preference mining). Finally, I study an efficient learning task to learn some parameters of a multicriteria decision method called MR-Sort by using support vector machines (SVM). This method rank alternatives in totally ordered classes.
Keywords : Lattice theory, graph theory, complexity, (Graph) preference learning and mining, conditional preference networks, noisy data.
My CV (in french) is available here.
My personal website is available at the address fabien.labernia.fr.