A0729
Title: Smooth backfitting of proportional hazards with multiplicative components
Authors: Munir Hiabu - University of Copenhagen (Denmark) [presenting]
Abstract: A proportional hazard model is proposed, where we assume an underlying conditional hazard with multiplicative components. No structural assumption is made on the components. The model is a generalization of the Cox proportional hazard model where components are assumed to be log-linear. We will fit the model via a smooth backfitting approach. Smooth backfitting has proven to have a number of theoretical and practical advantages in structured regression. It projects the data down onto the structured space of interest providing a direct link between data and estimator. Those ideas will be brought to the field of survival analysis. Asymptotic theory for the proposed estimator will be developed. In a comprehensive simulation study, we show that our smooth backfitting estimator successfully circumvents the curse of dimensionality and outperforms existing estimators. This is especially the case in difficult situations with higher dimensions and/or high correlation where other estimators tend to break down.
