A warped failure time model for human mortality
In: Eilers, P. H. C. (Ed.): Proceedings of the 23rd International Workshop on Statistical Modelling, Utrecht 7 - 11 July, 2008, 149–154
Utrecht, Koninklijke Nederlandse Akademie van Wetenschappen (2008)
We present an extension of the accelerated failure time model for comparison of density functions. With this time model one can study how the time axis would have to be transformed so that one density distribution conforms to another. Parametric and non-parametric estimates from actual data can be used as target distributions. The only assumption we make about the warping function is the smoothness and we represent it by a linear combination of B-splines. To estimate the warping function we use a penalized Poisson likelihood approach. The optimal smoothing parameter is found by minimizing the BIC. Actual demographic applications are presented.