For the scientific and you can epidemiological scientific studies, persisted predictors are often discretized towards the categorical details to possess category out-of customers. In the event the dating anywhere between an ongoing predictor and you may log relative potential risks is You-shaped from inside the emergency research, there is deficiencies in a satisfying option to pick max cut-items to discretize new continuing predictor. Within research, we propose a ed max equal-Hours way of discretize a continuous adjustable who has a good You-formed experience of log relative dangers within the success data.
An element of the concept of the optimal equal-Time system is to obtain a couple of optimum reduce-points that possess equal record relative hazard values and cause Cox models that have lowest AIC worth. Leer más