A generalized -approach for a kernel estimator of conditional quantile with functional regressors: Consistency and asymptotic normality

Ali Laksaci , Mohamed Lemdani , Elias Ould-Saïd (1)

1 LMPA - Laboratoire de Mathématiques Pures et Appliquées Joseph Liouville

A kernel estimator of the conditional quantile is defined for a scalar response variable given a covariate taking values in a semi-metric space. The approach generalizes the median's -norm estimator. The almost complete consistency and asymptotic normality are stated.

Mots clés: Primary; secondary

Domaines: Mathématiques [math]

Ali Laksaci, Mohamed Lemdani, Elias Ould-Saïd. A generalized -approach for a kernel estimator of conditional quantile with functional regressors: Consistency and asymptotic normality. Statistics and Probability Letters, 2009, 79 (8), pp.1065-1073. ⟨10.1016/j.spl.2008.12.016⟩. ⟨hal-00516884⟩