On the relationship between α connections and the asymptotic properties of predictive distributions
Corcuera, José M. ; Giummolè, Federica
Bernoulli, Tome 5 (1999) no. 6, p. 163-176 / Harvested from Project Euclid
In a recent paper, Komaki studied the second-order asymptotic properties of predictive distributions, using the Kullback-Leibler divergence as a loss function. He showed that estimative distributions with asymptotically efficient estimators can be improved by predictive distributions that do not belong to the model. The model is assumed to be a multidimensional curved exponential family. In this paper we generalize the result assuming as a loss function any f divergence. A relationship arises between α connections and optimal predictive distributions. In particular, using an α divergence to measure the goodness of a predictive distribution, the optimal shift of the estimate distribution is related to α-covariant derivatives. The expression that we obtain for the asymptotic risk is also useful to study the higher-order asymptotic properties of an estimator, in the mentioned class of loss functions.
Publié le : 1999-02-14
Classification:  curved exponential family,  differential geometry,  f divergences,  predictive distributions,  second-order asymptotic theory,  α connections,  α embedding curvature
@article{1173707099,
     author = {Corcuera, Jos\'e M. and Giummol\`e, Federica},
     title = {On the relationship between $\alpha$ connections and the asymptotic properties of predictive distributions},
     journal = {Bernoulli},
     volume = {5},
     number = {6},
     year = {1999},
     pages = { 163-176},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1173707099}
}
Corcuera, José M.; Giummolè, Federica. On the relationship between α connections and the asymptotic properties of predictive distributions. Bernoulli, Tome 5 (1999) no. 6, pp.  163-176. http://gdmltest.u-ga.fr/item/1173707099/