Extension of model-based classification for binary data when training and test populations differ
Résumé
Standard discriminant analysis supposes that both the training sample and the test sample are issued from the same population. When these samples arise from populations differing from their descriptive parameters, a generalization of discriminant analysis consists in adapting the classification rule related to the training population to another rule related to the test population, by estimating a link between both populations. This paper extends an existing work available in a multinormal context to the case of binary data. To raise the major challenge which consists in defining a link between the two binary populations, it is supposed that binary data result from the discretization of latent Gaussian data. Estimation method and robustness study are presented, and two applications in a biological context illustrate this work.
Origine | Fichiers produits par l'(les) auteur(s) |
---|