Packages
lineagefreq
Lineage Frequency Dynamics from Genomic Surveillance Counts
Version: 0.2.0 (CRAN) / 0.6.0 (dev)
Multinomial logistic regression for pathogen variant frequency dynamics, growth advantage estimation, and short-term forecasting with rolling-origin backtesting. Five engines (frequentist and Bayesian) behind a unified interface; ships with real CDC SARS-CoV-2 data.
Links: CRAN | GitHub | Documentation
Install:
install.packages("lineagefreq")Core functions:
| Function | Description |
|---|---|
fit_model() |
Fit a lineage frequency model using any of five engines |
growth_advantage() |
Estimate relative growth advantages from a fitted model |
forecast() |
Generate short-term frequency forecasts |
backtest() |
Rolling-origin cross-validation for forecast evaluation |
score_forecasts() |
Score forecast accuracy with proper scoring rules |
survinger
Design-Adjusted Inference for Pathogen Lineage Surveillance
Version: 0.1.1
Design-adjusted prevalence estimation for genomic surveillance under unequal sequencing rates, using Horvitz–Thompson and Hajek estimators with right-truncation delay correction and Neyman-optimal resource allocation.
Links: CRAN | GitHub | Documentation
Install:
install.packages("survinger")Core functions:
| Function | Description |
|---|---|
surv_design() |
Define a surveillance sampling design |
surv_lineage_prevalence() |
Estimate lineage prevalence with design adjustment |
surv_optimize_allocation() |
Neyman-optimal allocation of sequencing resources |
surv_adjusted_prevalence() |
Adjusted prevalence with right-truncation correction |
clinicalfair
Algorithmic Fairness Assessment for Clinical Prediction Models
Version: 0.1.1
Model-agnostic fairness auditing for clinical prediction models. Computes group-stratified metrics with bootstrap CIs, flags four-fifths rule violations, and performs intersectional analysis and threshold-based mitigation.
Links: CRAN | GitHub | Documentation
Install:
install.packages("clinicalfair")Core functions:
| Function | Description |
|---|---|
fairness_data() |
Prepare data for fairness analysis |
fairness_metrics() |
Compute group-stratified performance metrics |
fairness_report() |
Generate a four-fifths rule compliance report |
threshold_optimize() |
Find group-specific thresholds to equalize metrics |
intersectional_fairness() |
Intersectional subgroup fairness analysis |
syntheticdata
Synthetic Clinical Data Generation and Privacy-Preserving Validation
Version: 0.1.1
Synthetic clinical data generation via Gaussian copula, bootstrap, and Laplace noise, with integrated distributional validation, privacy risk assessment, and downstream model fidelity testing.
Links: CRAN | GitHub | Documentation
Install:
install.packages("syntheticdata")Core functions:
| Function | Description |
|---|---|
synthesize() |
Generate synthetic data using copula, bootstrap, or noise methods |
validate_synthetic() |
Distributional validation of synthetic vs. original data |
privacy_risk() |
Membership inference and attribute disclosure risk assessment |
model_fidelity() |
Compare downstream model performance on real vs. synthetic data |
compare_methods() |
Benchmark multiple synthesis methods side by side |