Fig. 1From: Comparison of methods for early-readmission prediction in a high-dimensional heterogeneous covariates and time-to-event outcome frameworkIllustration of different situations when dealing with censored data that cannot be labeled when using a threshold ε. \(\delta _{i} = \mathbbm{1}_{\{T_{i} \leq C_{i}\}}\) is the censoring indicator which is equal to 1 if Yi is censored and 0 otherwise. In the binary outcome setting, patient 4 would be excludedBack to article page