This repository contains code for paper on random directed graph model for TB household contact study.
simulation.cpp: code of MCMC implemention for the model without powerful predictors, i.e., model 2 in the paper.simulation_powerful.cpp: code of MCMC implementation for the model with powerful predictors, i.e., model 1 in the paper.weight_comparison.cpp: code of computing DIC for the purpose of model comparison (with different weighting schemes).
out1.rds: DIC output for the first scenario, i.e., where extra-household transmission is predominant (accounts for about 90% of the total infections).out2.rds: DIC output for the second scenario, i.e., where extra-household transmission is stronger than household transmission (accounts for about 50%-60% of the total infections).out3.rds: DIC output for the third scenario, i.e., where household transmission is stronger than extra-household transmission (accounts for about 50%-60% of the total infections).out4.rds: DIC output for the fourth scenario, i.e., where household transmission is predominant (accounts for about 90% of the total infections).output.R: Proprocessing code.prediction.rds: Simulation results based on the Brazilian household contact study. Used as the final interpretable results given by the random directed graph model.si_hhc.rds: Posterior sample obtained based on the household contact study data only (without community controls). i = 1, 2, 3 or 4 signaling the scenario ID.si_hw.rds: 95% credible intervals obtained based on the household contact study data under optimal weighting schemes.si_whole.rds: Posterior sample obtained based on the whole data (with perfect community controls).simulation_prediction.R: The R code used to generateprediction.rds.unweight_est.rds: Posterior sample obtained based on the Brazilian household contact study without weights.weighted_est.rds: Posterior samples obtained based on the Brazilian household contact study under different weighting schemes.weight_DIC.rds: DIC for posterior samples inweighted_est.rds.
powersim.R: data simulation function for the model with powerful predictors, i.e., model 1 in the paper.simulated_data.R: data simulation function for the model without powerful predictors, i.e., model 2 in the paper.