Fig. 1From: Comparison of State-of-the-Art Neural Network Survival Models with the Pooled Cohort Equations for Cardiovascular Disease Risk PredictionFrameworks for neural network survival models. The frameworks of Deepsurv/Cox-nnet and Nnet-survival are shown in A, B respectively. In Fig. 1A, the Deepsurv and Cox-nnet model output \({h}_{w}\left({{\varvec{X}}}_{i}\right)\) which is used to replace the log risk \({{\varvec{X}}}_{i}^{T}{\varvec{\beta}}\) in the Cox model. In Fig. 1B, the output layers generate \({h}_{j}^{i}\) which is the hazard for individual i at time jBack to article page