Plot probability of infection

Plot probabilities of infection from compartmental model. Returns the probabilities and the plot.

plot_prob_infection(
  chain,
  nsamps,
  INCIDENCE_FUNC,
  solve_times,
  obs_dat = NULL,
  true_prob_infection = NULL,
  tshift = 0,
  smooth = FALSE
)

Arguments

chain	A dataframe containing the MCMC samples
nsamps	Number of samples
INCIDENCE_FUNC	A pointer to the Gaussian process model
solve_times	Vector indicating the time over which the model is solved
obs_dat	A dataframe containing observed Ct values and time of sample collection. NULL by default.
true_prob_infection	A dataframe from simulated data with two columns, one for time and the other is the true probability of infection. NULL by default.
tshift	Shift the solve times? Numeric, set to 0 by default
smooth	Smooth the model estimates for plotting? FALSE by default.

Value

Return a list containing three things: 1. A dataframe of model predictions containing time, probability of infection, and sample number; 2. A dataframe containing the maximum posterior probability of infection and time; 3. A ggplot showing the probabilities of infection

See also

Other plots: plot_distribution_fits(), predicted_distribution_fits()

Author

James Hay, jhay@hsph.harvard.edu

Examples

data(example_seir_incidence)
predictions <- plot_prob_infection(chain_comb, 
                                 nsamps=100, 
                                 INCIDENCE_FUNC=incidence_function,
                                 solve_times=0:max(ct_data_use$t),
                                 obs_dat=ct_data_use,
                                 true_prob_infection=example_seir_incidence)
#> Error in unique(chain$sampno) object 'chain_comb' not found
p_incidence_prediction <- predictions$plot + scale_x_continuous(limits=c(0,200))
#> Error in eval(expr, envir, enclos) object 'predictions' not found
p_incidence_prediction
#> Error in eval(expr, envir, enclos) object 'p_incidence_prediction' not found