sim_data: Simulate data
In Michael-Stevens-27/silverblaze: Geographic Profiling in R

Simulate data under a specified model

sim_data(
  sentinel_lon,
  sentinel_lat,
  sentinel_radius = 0.1,
  K = 3,
  source_weights = NULL,
  source_lon_min = -0.2,
  source_lon_max = 0,
  source_lat_min = 51.45,
  source_lat_max = 51.55,
  source_lon = NULL,
  source_lat = NULL,
  sigma_model = "single",
  sigma_mean = 1,
  sigma_var = 0.1,
  expected_popsize = 100,
  data_type = "counts",
  test_rate = 5,
  N = 150,
  dispersal_model = "normal"
)

`sentinel_lon`	vector giving longitudes of sentinel sites.
`sentinel_lat`	vector giving latitudes of sentinel sites.
`sentinel_radius`	observation radius of the sentinel site (km).
`K`	the number of sources.
`source_weights`	the proportion of events coming from each source
`source_lon_min`	minimum limit on source longitudes.
`source_lon_max`	maximum limit on source longitudes.
`source_lat_min`	minimum limit on source latitudes.
`source_lat_max`	maximum limit on source latitudes.
`source_lon`	manually define source longitude positions. If `NULL` then drawn uniformly from limits specified in `source_lon_min` and `source_lon_max`.
`source_lat`	manually define source latitude positions. If `NULL` then drawn uniformly from limits specified in `source_lat_min` and `source_lat_max`.
`sigma_model`	set as "single" to use the same dispersal distance for all sources, or "independent" to use an independently drawn dispersal distance for each source.
`sigma_mean`	the prior mean of the parameter sigma (km).
`sigma_var`	the prior variance of the parameter sigma (km). Set to zero to use a fixed distance.
`expected_popsize`	the expected total number of observations (observed and unobserved) in the study area.
`data_type`	what model we wish to simulate under - a poisson, binomial or vanilla finite mixture corresponding to "counts", "prevalence" or "point-pattern" respectively
`test_rate`	The rate of the Poisson distribution with which we draw the number of individuals tested at each sentinel site
`N`	the number of events to distributed under a point-pattern model
`dispersal_model`	distribute points via a "normal", "cauchy" or "laplace" model

# State the number of sources to be generated
K_sim <- 3
# Create some sentinel site locations
sentinal_lon <- seq(-0.2, 0.0, l=11)
sentinal_lat <- seq(51.45, 51.55, l=11)
sentinal_grid <- expand.grid(sentinal_lon, sentinal_lat)
names(sentinal_grid) <- c("longitude", "latitude")
# Set their sentinel radius (this constant times true sigma)
sentinel_radius <- 0.25
# sim count data under a Poisson model
sim1 <- sim_data(sentinal_grid$longitude,
                sentinal_grid$latitude,
                sigma_model = "single",
                sigma_mean = 1,
                sigma_var = 0.5,
                sentinel_radius = sentinel_radius,
                K = K_sim,
                expected_popsize = 300)