Greenwood Roadmap

Greenwood is built in dependency-ordered, individually shippable steps. This is the public capability roadmap.

Planned — Long Term

Advanced estimators for complex survival problems and specialized applications.

Additive Hazards & Cure Models

Alternative hazard structures and zero-inflated survival models.

  • Aalen additive model for additive (vs. proportional) hazard regression with constrained optimization to ensure non-negative hazards and proper survival functions
  • Mixture cure models for populations with long-term survivors
  • Non-parametric maximum likelihood estimation (NPMLE) for cure fractions
  • Goodness-of-fit tests and model comparison for cure models

Advanced Competing Risks & Multi-State

Extended methods for cause-specific and multi-state analyses.

  • Variance estimation for multi-state transition probabilities
  • Pseudo-observation approach for CIF and multi-state occupancy regression
  • Custom estimands via pseudo-observations framework

Frailty Models

Random-effects Cox models for correlated survival data.

  • Conditional and marginal predictions for known and new clusters

Advanced Performance Metrics

Discrimination and calibration assessment beyond point-in-time.

  • Integrated discrimination improvement (IDI) and net reclassification improvement (NRI)
  • Time-dependent Brier score refinements and sensitivity analyses
  • ipc_weights / CensoringDistributionEstimator: IPC weights as first-class utilities, enabling manual reweighting for metrics and models beyond concordance

Machine Learning Survival Estimators

Tree-based and ensemble survival models that output individual survival functions rather than scalar risk scores.

  • SurvivalTree: single survival tree using the log-rank split criterion, which is the foundation for ensemble methods and useful for interpretable non-parametric survival modeling
  • RandomSurvivalForest: ensemble of survival trees with bootstrap aggregation (the most widely used ML survival model, returns per-subject survival and cumulative hazard functions)
  • ExtraSurvivalTrees: extremely randomized variant of RSF with additional split randomization for faster training and lower variance on large datasets
  • GradientBoostingSurvivalAnalysis: component-wise and tree-based boosting for survival (strong predictive performance on structured tabular data)
  • IPCRidge: IPC-weighted ridge regression (a linear survival model that corrects for censoring bias via inverse-probability reweighting rather than the partial likelihood)

Platform & Interop

Performance and ecosystem integration toward 1.0.

  • Full backend matrix algebra with accelerated kernels (JAX/Numba) for ultra-large datasets (100k+ rows)
  • Finalized extension protocols and Narwhals dataframe backend completeness
  • Full interoperability with Great Summaries (tbl_survfit, tbl_regression)
  • Migration guides for users transitioning from R’s survival package

Feedback & Contributions

Have ideas for features not listed here? Open an issue with the enhancement label! Contributions to any planned item are welcome so check existing issues first to avoid duplication.

This roadmap is a living document. It is updated as features are added and new priorities emerge.