Indeed, if this is not done, the type-I error is not controlled, and then it is possible to obtain some false positive results. Models that have larger predicted R 2 values have better predictive ability. We are currently working on the generalization of these methods to proportional hazard models, which we will make available as soon as possible in the CPMCGLM package. Computation of the p -value of the minimum of score tests in the generalized linear model, application to multiple coding. We can use these values to calculate the X 2 statistic of the model:.
nest...