Estimation locale : compromis régression-variance

Olivier Barrière (1) , Jacques Lévy Véhel (2)

1 IRCCyN - Institut de Recherche en Communications et en Cybernétique de Nantes
2 APIS - Analysis of irregular Processes, Images and Signal,applications in biology and medicine

We present a method allowing under some hypothesis to decrease the variance of an optimal estimator without causing damage to its risk, using many estimators jointly, thanks to a very basic technique leading to a hybrid estimator that takes advantage of the best of them. This method can be applied in various fields, typically on estimators that were obtained using linear regressions.

Domaines: Probabilités [math.PR]

Olivier Barrière, Jacques Lévy Véhel. Estimation locale : compromis régression-variance. Colloque GRETSI, Sep 2007, Troyes, France. pp.593-596. ⟨hal-00539212⟩