NEWS
bayespk 0.1.0
- First release: 1- and 2-compartment linear models (IV bolus, infusion, first-order oral), closed-form solutions, log-normal between-subject variability (optionally correlated), covariate formulas, four residual error models, weakly informative default priors, diagnostics with warnings, predictions for new regimens, prior sensitivity.
pk_model(ka_faster = TRUE) constrains the typical KA above the disposition rate to remove flip-flop modes (off by default).
- Random effects are sampled as a sample mean plus Helmert deviations, which removes the correlation between typical values and the mean random effect.