Dear Colleagues

I came across this paper: "Balanced Random Survival Forests for Extremely Unbalanced, Right Censored Data" I have attached it to this discussion.

I am not sure how they did the balancing and the SMOTE in this paper.

My understanding is that they divided the data into training and test dataset.

Then, they did oversampling for the events, to make the dataset equal in terms of events.

After that they ran the random survival forest.

In order to do that, they added the " time" to the features, and kept the "event" as the outcome , then they did the SMOTE. After that they moved the "time" out of the features and then ran the random survival forest

Is my understanding correct? If not, can you please explain to me how they did the SMOTE for the right censored data? and how to do it?