It is necessary to check the predictions for new individuals (or even the training data individuals) are correct.
As a first step, for the new individuals, I should fit the model twice -- once with them included, and once with them excluded (just for a couple of individuals) -- and compare the b pars, and the b_new pars, from the two fits. As well as the posterior_survfit predictions.
It is necessary to check the predictions for new individuals (or even the training data individuals) are correct.
As a first step, for the new individuals, I should fit the model twice -- once with them included, and once with them excluded (just for a couple of individuals) -- and compare the b pars, and the b_new pars, from the two fits. As well as the posterior_survfit predictions.