Running a likelihood profile on a FIMS model • FIMSdiags

library(FIMSdiags)
library(FIMS)
#devtools::load_all()
packageVersion("FIMS")
#> [1] '0.7.1'

clear()

Set up FIMS model

Following the minimal demo vignette, we will set up a FIMS model and run it.


# Load sample data
data("data1")
# Prepare data for FIMS model
data_4_model <- FIMSFrame(data1)

# Create parameters
parameters <- data_4_model |>
  create_default_configurations() |>
  create_default_parameters(data = data_4_model)

# Run the  model with optimization
base_model <- parameters |>
  initialize_fims(data = data_4_model) |>
  fit_fims(optimize = TRUE)
#> ✔ Starting optimization ...
#> ℹ Restarting optimizer 3 times to improve gradient.
#> ℹ Maximum gradient went from 0.02469 to 0.00037 after 3 steps.
#> ✔ Finished optimization
#> ✔ Finished sdreport
#> ℹ FIMS model version: 0.7.1
#> ℹ Total run time was 10.08182 seconds
#> ℹ Number of parameters: fixed_effects=77, random_effects=0, and total=77
#> ℹ Maximum gradient= 0.00037
#> ℹ Negative log likelihood (NLL):
#> • Marginal NLL= 3240.1261
#> • Total NLL= 3240.1261
#> ℹ Terminal SB= 1803.27712

# Clear memory post-run
clear()

Run likelihood profile

Once we have a model that is fit, we can run a likelihood analysis to see the impact of each data type on a given parameter. In this example, we will test the impact on R0 but any value can be tested by changing the parameter_name argument. The parameter name must match what is

like_fit <- run_fims_likelihood(
  model = base_model,
  parameters = parameters,
  parameter_name = "log_rzero", 
  data = data1,
  min = -1,
  max = 1,
  length = 10
)

plot_likelihood(like_fit)


clear()