How thehanlab/dynamiclm works
dynamicLM is an R package for dynamic survival-risk prediction: it prepares cohort data, fits landmark supermodels, predicts updated risks, and evaluates or visualizes results. The README documents the principal workflow and capabilities, but no source excerpts were provided; edges therefore reflect documented relationships, not inspected function calls.
Main entry points: predict.R Risk prediction · stack_data.R Stacked data · score.R Prediction scoring · get_pbc_long.R PBC loader
Components 17
Data preparation
- PBC loader
R/get_pbc_long.R - Stacked data
R/stack_data.R - Time interactions
R/add_interactions.R - Model data
R/get_lm_data.R
Landmark modeling
- Landmark model
R/dynamic_lm.R - Penalized model
R/pen_lm.R - Cross-validation
R/cv.pen_lm.R - Model helpers
R/dynamic_lm_helper.R
Risk prediction
- Risk prediction
R/predict.R - Risk scores
R/riskScore.R
Evaluation and plots
- Prediction scoring
R/score.R - Performance metrics
R/summary_metric.R - Calibration plots
R/calplot.R - Risk trajectories
R/plotrisk.R - Model plots
R/plot.R
Other components
- Analyst
- Survival data
Connections 12
- Analyst → PBC loader: loads data
README.md - Survival data → PBC loader: combines data
README.md - Analyst → Stacked data: prepares cohort
README.md - Stacked data → Landmark model: fits on
README.md - Stacked data → Penalized model: fits on
README.md - Analyst → Cross-validation: validates model
README.md - Analyst → Risk prediction: predicts risk
README.md - Risk prediction → Prediction scoring: evaluates predictions
README.md - Risk prediction → Calibration plots: checks calibration
README.md - Risk prediction → Risk trajectories: visualizes trajectories
README.md - Risk prediction → Model plots: visualizes model
README.md - Prediction scoring → Performance metrics: reports metrics
README.md