Many methods can be used to reduce a discrete survival distribution prediction (i.e. matrix) to a relative risk / ranking prediction. Here we define the predicted relative risk as the sum of the predicted cumulative hazard function - which can be loosely interpreted as the expected number of deaths for patients with similar characteristics.
surv_to_risk(x)(matrix())
TxN survival matrix prediction where T is number
of time-points and N is number of predicted observations. Colum names
correspond to predicted time-points and should therefore be coercable to
numeric and increasing. Entries are survival predictions and should
be (non-strictly) decreasing in each row.
Sonabend, R., Bender, A., & Vollmer, S. (2021). Evaluation of survival distribution predictions with discrimination measures. http://arxiv.org/abs/2112.04828.