Skip to content

Latest commit

 

History

History
119 lines (107 loc) · 4.91 KB

Quick_start_overview.md

File metadata and controls

119 lines (107 loc) · 4.91 KB
Raw

Quick start guide to running EvoNet

Download/install/load EvoNet

if (!require("devtools")) install.packages("devtools")
install_github("EvoNetHIV/TestRepo",subdir="pkg")
library(evonet)

Load default parameters.

primary_parameters  <- input_parameters_primary()
cd4_data            <- input_parameters_cd4_data()

Combine individual parameters into single list. Parameter list can be viewed by entering 'evoparams' in RStudio console.

evoparams <- c(primary_parameters, cd4_data)

Change default parameters. In this example, we change initial population size to 200 (from 100); number of initially infected agents to 40 (from 20); and the duration of the model to 20 years.

evoparams$initial_pop      = 200
evoparams$initial_infected = 40
evoparams$n_steps = 365*20

Calculate derived parameters (parameters that are functions of other parameters)

evoparams  <- input_parameters_derived(evoparams)

Convert initial parameter list into EpiModel parameter list (so EpiModel recognizes it as a parameter list)

evoparams <- do.call(EpiModel::param.net,evoparams)

Check to make sure input parameters are valid (error returned if not)

input_parameters_qaqc(evoparams)

Create initial network (as a function of input parameters)

nw <- setup_initialize_network(evoparams)

Create list of arguments for EpiModel's network estimation function.

netest_arg_list <- list(
  nw            =  nw,
  formation     =  as.formula(evoparams$nw_form_terms),
  target.stats  =  evoparams$target_stats,
  coef.form     =  evoparams$nw_coef_form,
  constraints   =  as.formula(evoparams$nw_constraints),
  verbose       =  FALSE,
  coef.diss     =  dissolution_coefs( dissolution =  as.formula(evoparams$dissolution),
                                      duration    =  evoparams$relation_dur,
                                      d.rate      =  3e-05) )

Estimate network (i.e., create desired network structure and dynamics)

estimated_nw <- do.call(EpiModel::netest, netest_arg_list)

Create vector of infection status (0/1) as an epimodel object for initial population

infected_list <- EpiModel::init.net(i.num=evoparams$initial_infected,
                                    status.rand = FALSE)

Create list with modules/functions to simulate epidemic with desired dynamics.

evo_module_list<- list(
  "initialize.FUN"     = initialize_module,
  "plot_nw.FUN"        = plot_network_fxn,  
  "aging.FUN"          = vital_aging_module,
  "testing.FUN"        = social_testing_diagnosis_module,
  "treatment.FUN"      = social_treatment_module,
  "update_vl.FUN"      = viral_update_gamma,
  "update_cd4.FUN"     = viral_update_cd4_daily, 
  "coital_acts.FUN"    = social_coital_acts_module,
  "trans.FUN"          = transmission_main_module,
  "trans_book.FUN"     = transmission_bookkeeping_module,
  "trans_cd4.FUN"      = transmission_cd4_module,
  "deaths.FUN"         = vital_deaths_module,
  "births.FUN"         = vital_births_module,
  "summary.FUN"        = summary_module,
  "resim_nets.FUN"     = EpiModel::resim_nets,
  "verbose.FUN"        = NULL)

Create an EpiModel "control" object which contains both the input parameters and the modules

evocontrol <- setup_epimodel_control_object(evonet_params = evoparams,
                                            module_list   = evo_module_list)

Run the simulation using EpiMode's 'netsim' function

  evomodel  <- EpiModel::netsim(x = estimated_nw,
                                param = evoparams,
                                init = infected_list,
                                control = evocontrol)

Initial network plot will be plotted by default at simulation start (colored circles represent infected agents). (Day 2 is first day of simulation as model initialization is considered day 1.)

Save model output. Default location is current working directory.

save(evomodel,
     file = file.path(evoparams$output_path,"evomodel.RData"))

Create default output plots summarizing model run. Plots printed to screen and saved as pdf file (default name: "popsumm_figures.pdf") to working directory.

plots_popsumm(evomodel,outpath=evoparams$output_path,
              name=NULL,nw_stats=TRUE,max_points_rep=100,
              evoparams$popsumm_frequency)

Example of two default output plots