Skip to Main content Skip to Navigation

Estimation non-paramétrique de données censurées dans un cadre multi-états

Abstract : This thesis deals with competing risks and recurrent events.
In the competing risks model, the
interest is centered on the cumulative incidence functions. These functions correspond to the probability
that a given kind of event happens before a given time. These functions are estimated by means
of the nonparametric estimator of Aalen-Johansen. Strong approximations by Gaussian processes,
laws of the iterated logarithm and weak convergence results for processes based on the
Aalen-Johansen estimator are established. Asymptotic confidence bands are constructed. Moreover,
a generalization of the Koziol-Green model is considered.
In the recurrent events model, conditional cumulative incidence functions are estimated
nonparametrically. The proposed estimators are strongly consistant. The finite distance behavior is
investigated by means of Monte-Carlo simulations and illustrated on real data.
Document type :
Complete list of metadatas
Contributor : Ségolen Geffray <>
Submitted on : Sunday, March 25, 2007 - 10:24:22 AM
Last modification on : Friday, November 20, 2020 - 8:55:34 AM
Long-term archiving on: : Tuesday, April 6, 2010 - 10:57:25 PM


  • HAL Id : tel-00138280, version 1


Ségolen Geffray. Estimation non-paramétrique de données censurées dans un cadre multi-états. Mathematics [math]. Université Pierre et Marie Curie - Paris VI, 2006. English. ⟨tel-00138280⟩



Record views


Files downloads