Mortality modeling and regression with matrix distributions

In this paper we investigate the flexibility of matrix distributions for modeling mortality. Starting from a simple Gompertz law, show how introduction matrix-valued parameters via inhomogeneous phase-type can lead to reasonably accurate and relatively parsimonious models mortality curves across entire lifespan. A particular feature proposed model framework is that it allows more direct interpretation implied underlying aging process than some previous approaches. Subsequently, towards applications approach multi-population modeling, introduce regression concept proportional intensities, which are flexible hazard models, two classes asymptotically equivalent. We illustrate be estimated data by providing an adapted EM algorithm likelihood increases at each iteration. The practical feasibility competitiveness approach, including right-censored case, illustrated several sets survival data.