Modèle de durée pour données multivoie

Abstract : In this paper, we present a multiway extension of Cox proportional hazards model (Multiway Cox). Multi-way data refers to the case where several modalities of the same variable have been collected for each individual. The approach consists in imposing a tensor-like structure to the weight coefficients echoing the structure of the multi-way data set. This approach reduces the number of coefficients to estimate and limits overfitting issues. The proposed procedure is evaluated and validated on simulated data.
Complete list of metadatas
Contributor : Laurent Le Brusquet <>
Submitted on : Thursday, February 23, 2017 - 1:30:56 PM
Last modification on : Friday, October 19, 2018 - 9:44:02 PM
Long-term archiving on : Wednesday, May 24, 2017 - 1:50:43 PM


JdS 2017 - Modele de durée & ...
Files produced by the author(s)


  • HAL Id : hal-01474989, version 1


Alfred Baroulier, Mehdi Douch, Paul Messinesi, Laurent Le Brusquet, Gisela Lechuga, et al.. Modèle de durée pour données multivoie. 49émes Journées de Statistique, May 2017, Avignon, France. ⟨hal-01474989⟩



Record views


Files downloads