This function sets up and runs a particle MCMC that uses Dust, Odin and MCState
Usage
run_pmcmc(
data_raw = NULL,
data_raw_pg = NULL,
data_raw_mg = NULL,
init_EIR = 10,
target_prev = NULL,
target_prev_group = "u5",
n_particles = 200,
proposal_matrix = matrix(1),
max_param = 125,
prop_treated = 0.4,
n_steps = 500,
n_threads = 4,
n_chains = 1,
n_workers = 1,
state_check = 0,
country = NULL,
admin_unit = NULL,
seasonality_on = TRUE,
seasonality_check = FALSE,
check_flexibility = FALSE,
seed = 1L,
start_pf_time = 30 * 12,
particle_tune = FALSE,
comparison = "u5",
initial = "informed"
)Arguments
- data_raw
Time series data to fit model
- data_raw_pg
Time series of primigravidae ANC data to fit model
- data_raw_mg
Time series of multigravidae ANC data to fit model
- init_EIR
A single value or a dataframe with two columns (time and EIR) to specify historical malaria transmission levels before data collection began.
- target_prev
Return an initial EIR value (from the equilibrium solution), given a target prevalence in under 5yos
- target_prev_group
Age group used for target prevalence
- n_particles
Number of particles to be used in pMCMC (default = 200)
- proposal_matrix
Proposal matrix for MCMC parameters
- max_param
Ceiling for proposed stochastic parameter (either EIR or betaa) values (default = 1000)
- prop_treated
Proportion of clinical cases that receive effective treatment (default = 40%)
- n_steps
Number of MCMC steps in a single chain (default = 500)
- n_threads
Number of processing threads (default = 4)
- n_chains
Number of chains (default = 1)
- n_workers
Number of workers (default = 4)
- state_check
If state_check = TRUE, returns expected deriv values which should equal 0 and sets stochastic model to have EIR constant at init_EIR If state_check = TRUE and seasonality_on = 1, then the deterministic seasonal model is still run, but theta2 is forced to 1, forcing a constant seasonality profile If state_check = FALSE, no values are printed
- country
Name of country (needed if using seasonality model)
- admin_unit
Name of administrative unit (needed if using seasonality model)
- seasonality_on
Toggle seasonality model run before observed time period (default = 1)
- seasonality_check
Toggle saving values of seasonality equilibrium (default = 1)
- check_flexibility
Toggle saving values of flexibility period
- seed
Allows user to specify a seed (default = 1L)
- start_pf_time
Number of days before first observation that particle filter will start (default = 30)
- particle_tune
Logical to determine if tuning the number of particles should be performed.
- comparison
The comparison function to be used. Either 'u5' which equates the observed prevalence to prevalence under 5 years old in the model or 'pgmg' which calculates prevalence in primigravid and multigravid pregnant women for comparison with observed ANC data. c('u5','pg','sg','mg','pgmg','pgsg','ancall') If in a format XtoY, where X and Y are two numbers, will compare to general population between those two ages.
- initial
Is the initial equilibrium state informed by the user ('user-informed') or by the observed data ('fitted')?