Goodness-of-fit test for the AFT model based on martingale residuals

The Accelerated Failure Time model presents a way to easily describe survival regression data. It is assumed, that each observed unit ages internally faster or slower, depending on the covariate values. To use the model properly, we want to check if observed data fit the model assumptions. In present work we introduce a goodness-of-fit statistics based on modern martingale theory. On simulated data we try to estimate empirical properties of the test for various situations.