L'analyse implicative bayésienne : une méthode pour l'étude des dépendances orientées. 1.Données binaires

摘要: Does success to test A imply, approximately, B ? Among the descriptive indices that were proposed measure such oriented dependencies, we focus on Loevinger's H which has a simple expression in terms of association rates between modalities. From this index, define notions quasi-implication, quasi-equivalence and quasi-independence for 2 by contingency table. However, corresponding existing inferential frequentist methods, because they are based asymptotic considerations, do not provide valid statements, neither small datasets, nor, paradoxically, cases implication is descriptively almost verified. The Bayesian approach inference avoids these difficulties provides inductive statements about size quasi-implications other types "quasi-" relations. Implicative Analysis, composed two aspects (descriptive inductive), illustrated an experiment from Developmental Psychology number construction children.