A classification method for binary predictors combining similarity measures and mixture models

Seydou Nourou Sylla (1, 2, 3, 4) , Stéphane Girard (4, 1) , Abdou Ka Diongue (3, 4) , Aldiouma Diallo (2, 4) , Cheikh Sokhna (2)

1 MISTIS - Modelling and Inference of Complex and Structured Stochastic Systems
2 URMITE - Unité de Recherche sur les Maladies Infectieuses Tropicales Emergentes
3 LERSTAD - laboratoire d'Etudes et de recherches en Statistiques et Développement
4 LIRIMA - Laboratoire International de Recherche en Informatique et Mathématiques Appliquées

In this paper, a new supervised classification method dedicated to binary predictors is proposed. Its originality is to combine a model-based classification rule with similarity measures thanks to the introduction of new family of exponential kernels. Some links are established between existing similarity measures when applied to binary predictors. A new family of measures is also introduced to unify some of the existing literature.The performance of the new classification method is illustrated on two real datasets (verbal autopsy data and handwritten digit data) using 76 similarity measures.

Mots clés: similarity measure; Mixture model; binary data; kernel method

Domaines: Statistiques [math.ST]

Seydou Nourou Sylla, Stéphane Girard, Abdou Ka Diongue, Aldiouma Diallo, Cheikh Sokhna. A classification method for binary predictors combining similarity measures and mixture models. Dependence Modeling, 2015, 3, pp.240-255. ⟨10.1515/demo-2015-0017⟩. ⟨hal-01158043v5⟩