Divergence measures and logistic regression models

  1. Pardo Llorente, Leandro
Livre:
VIII Journées Zaragoza-Pau de Mathématiques Appliquées et de Statistiques
  1. Palacios Latasa, Manuel Pedro (coord.)
  2. Trujillo, David (coord.)
  3. Torrens Iñigo, Juan José (coord.)
  4. Madaune-Tort, Monique (coord.)
  5. López de Silanes Busto, María Cruz (coord.)
  6. Sanz Sáiz, Gerardo (coord.)

Éditorial: Prensas de la Universidad de Zaragoza ; Universidad de Zaragoza

ISBN: 84-7733-720-9

Année de publication: 2003

Pages: 415-423

Congreso: Jornadas Zaragoza-Pau de Matemática Aplicada y Estadística (8. 2003. Jaca)

Type: Communication dans un congrès

Résumé

In this paper we present a review of some results about inference based on f-divergence measures, under assumptions of logistic regression model . The minimum f-divergence estimator, which is seen to be a generalization of the maximum likelihood estimator is considered. This estimator is used in a f-divergence measure which is the basis of new statistics for solving some important problems regarding logistic regression models: fitting the logistic regression model, residuals and dimensional reduction. Finally, an extension is presented when we consider a multinomial response instead a binary response.