Goodness-of-fit inference for the Cox-Aalen additive-multiplicative regression model

The Cox-Aalen additive-multiplicative intensity model of Scheike and Zhang (Scand. J. Statist. 29, 2002) is considered. We study goodness-of-fit tests based on the stratified martingale residual process. Asymptotic distribution of the process is derived and the Kolmogorov-Smirnov type test is constructed. Several ways of overcoming the problem of complexity of the limiting distribution are discussed. The results are accompanied by a small Monte Carlo study.