In a proportional hazards model, if groups of observations are expected to have different underlying hazard rates, the model can be stratified using a category column. A common example is stratifying medical data by sex, so that a separate hazard rate is estimated for males and females.
To stratify the model using a category column:
For purposes of prediction, individual strata may be selected in the Predict view using the pop-up button below the survival curve in the bottom-left corner.
See also: