Explainable Machine Learning in Survival Analysis
Calculate Brier score
Compute the Harrell's Concordance index
Calculate Cumulative/Dynamic AUC
Transform Cumulative Hazard to Survival
A model-agnostic explainer for survival models
Extract Local SurvSHAP(t) from Global SurvSHAP(t)
Calculate integrated Brier score
Calculate integrated C/D AUC
Adapt mlr3proba measures for use with survex
Calculate integrated metrics based on time-dependent metrics.
Calculate the Concordance index loss
Calculate Cumulative/Dynamic AUC loss
Calculate integrated C/D AUC loss
Dataset Level Model Diagnostics
Dataset Level Variable Importance for Survival Models
Dataset Level Performance Measures
Dataset Level Variable Profile as Partial Dependence Explanations for ...
Dataset Level 2-Dimensional Variable Profile for Survival Models
Global SHAP Values
Plot Aggregated SurvSHAP(t) Explanations for Survival Models
Plot Model Diagnostics for Survival Models
Plot Model Parts for Survival Models
Plot Model Performance for Survival Models
Plot 2-Dimensional Model Profile for Survival Models
Plot Model Profile for Survival Models
Plot Predict Parts for Survival Models
Plot Predict Profile for Survival Models
Plot Permutational Feature Importance for Survival Models
Plot SurvLIME Explanations for Survival Models
Plot Model Performance Metrics for Survival Models
Plot ROC Curves for Survival Models
Plot SurvSHAP(t) Explanations for Survival Models
Model Predictions for Survival Models
Instance Level Parts of Survival Model Predictions
Instance Level Profile as Ceteris Paribus for Survival Models
Generate Risk Prediction based on the Survival Function
Helper functions for predict_profile.R
Helper functions for model_parts.R
Helper functions for model_parts.R
Helper functions for predict_parts.R
Extract additional information from the model
Helper functions for model_performance.R
Helper functions for predict_parts.R
Transform Survival to Cumulative Hazard
Default Theme for survex plots
Transform Fixed Point Prediction into a Stepfunction
Survival analysis models are commonly used in medicine and other areas. Many of them are too complex to be interpreted by human. Exploration and explanation is needed, but standard methods do not give a broad enough picture. 'survex' provides easy-to-apply methods for explaining survival models, both complex black-boxes and simpler statistical models. They include methods specific to survival analysis such as SurvSHAP(t) introduced in Krzyzinski et al., (2023) <doi:10.1016/j.knosys.2022.110234>, SurvLIME described in Kovalev et al., (2020) <doi:10.1016/j.knosys.2020.106164> as well as extensions of existing ones described in Biecek et al., (2021) <doi:10.1201/9780429027192>.