README.md

Code to reproduce paper

Bayes-raking: Bayesian Finite Population Inference with Known Margins

layouts

1. Data:

• acs_simu.rds: Data used for simulation, which mimic ACS
• lsw.rds: Data used for real data analysis. EXCLUDED IN THIS REPO

2. Scripts:

   • utils.R some functions used in simulation.
     • loading_matrix used to generate matrix $L$, describe in section 2.1 eq(5)
     • margin_vector used to generate vector $\overrightarrow{N}_{..}$
     • bayes_sample_contingency a small wrapper to genenar a sample contingency table information for bayes raking. Used similar grammar as R package survey::rake
     • NO NEED TO READ

   a. Section 3.1

   • sec3_1_simu.R TIME CONSUMING
   • sec3_1_one_realization.R With detail step how to generate loading matrix L.
   • stan/bayes_raking.stan Bayes raking + Binary outcome

   b. Section 3.2

   • sec3_2_simu.R TIME CONSUMING
   • sec3_2_one_realization.R With detail step how to generate loading matrix L.
   • bayes_raking_32.stan Bayes raking + continue outcome
   • basis_32.stan Basis + continue outcome
   • projection_32.stan Projection + continue outcome

   c. Section 4 (Real data, require Data/lsw.rds)

   • sec4_lsw_study.R !!REQUIRE lsw.rds required
   • bayes_raking_horseshoe.stan