Skip to contents

Visually compare prior and posterior of a model parameter

Source: R/sewer_plot.R
plot_prior_posterior.Rd

Visually compare prior and posterior of a model parameter

Usage

plot_prior_posterior(result, param_name)

Arguments

result

Results object returned by EpiSewer() after model fitting. In contrast to other plotting functions, this cannot be a list of multiple result objects, because prior-posterior plots of multiple results simultaneously are currently not supported.

param_name

Name of the single parameter to be plotted. This can either be the raw name of the parameter in the stan model, or it can be the semantic name from the single parameter dictionary (see details). Note that this must be a single parameter, it cannot be a parameter array. Also, only original parameters (not transformed parameters) for which a prior can be specified in EpiSewer are supported.

Value

A plot showing the density of the prior (grey) and posterior (blue) for the respective parameter. Can be further manipulated using ggplot2 functions to adjust themes and scales, and to add further geoms.

Details

The following parameters can be visualized (if in the model):

  • measurement_noise_cv (nu_upsilon_a): Coefficient of variation (measurement noise)

  • dPCR_total_partitions (nu_upsilon_b_mu): Average total number of partitions in dPCR

  • dPCR_partition_variation (nu_upsilon_b_cv): Partition number variation in dPCR

  • dPCR_conversion_factor (nu_upsilon_c): Conversion factor in dPCR

  • pre_replicate_cv (nu_psi): Coefficient of variation (pre-PCR noise)

  • load_variation_cv (nu_zeta): Individual-level coefficient of load variation

  • infection_overdispersion (I_xi): Overdispersion of infections

  • seeding_intercept (iota_log_seed_intercept): Initial number of infections