survdnn_losses: Loss Functions for survdnn Models

In survdnn: Deep Neural Networks for Survival Analysis with R 'torch'

Loss Functions for survdnn Models

Description

These functions define various loss functions used internally by 'survdnn()' for training deep neural networks on right-censored survival data.

Usage

cox_loss(pred, true)

cox_l2_loss(pred, true, lambda = 0.001)

aft_loss(pred, true, sigma = 1, aft_loc = 0, eps = 1e-12)

coxtime_loss(pred, true)

Arguments

pred	A torch tensor of model predictions. Its interpretation depends on the loss function: loss = "cox" or "cox_l2": linear predictors (log hazard ratios). loss = "aft": predicted log survival times. loss = "coxtime": predicted time-dependent risk scores.
true	A tensor with two columns: observed time and status (1 = event, 0 = censored).
lambda	Regularization parameter for 'cox_l2_loss' (default: '1e-3').
sigma	Positive numeric scale parameter for the log-normal AFT model (default: '1'). In 'survdnn()', a learnable global scale can be used via 'survdnn__aft_lognormal_nll_factory()'.
aft_loc	Numeric scalar location offset for the AFT model on the log-time scale. When non-zero, the model is trained on centered log-times 'log(time) - aft_loc' for better numerical stability. Prediction should add this offset back: 'mu = mu_resid + aft_loc'.
eps	Small constant for numerical stability (default: '1e-12').

Value

A scalar 'torch_tensor' representing the loss value.

Supported Losses

- **Cox partial likelihood loss** ('cox_loss'): Negative partial log-likelihood used in proportional hazards modeling. - **L2-penalized Cox loss** ('cox_l2_loss'): Adds L2 regularization to the Cox loss. - **Accelerated Failure Time (AFT) loss** ('aft_loss'): Log-normal AFT **censored negative log-likelihood** (uses both events and censored observations). - **CoxTime loss** ('coxtime_loss'): Placeholder (see details). A correct CoxTime loss requires access to the network and the full input tensor.

Examples

# Used internally by survdnn()

survdnn documentation built on Jan. 8, 2026, 9:07 a.m.