Extended Data Fig. 7: Comparison of SHAP values and Cox proportional hazards coefficients.
From: Learning the natural history of human disease with generative transformers
Shown are analyses for 10 selected diseases, as stated in the titles. SHAP values (x-axis) are estimated by averaging individual values from different trajectories. Cox proportional hazard coefficients (y-axis) are estimated using a proportional hazards model with parameter regularization, resulting in a high number of zero coefficients. The non-zero Cox coefficients and SHAP show a high correlation.
