OTIS a01 Ruhela formulations (full DLRM) — morie_otis_analyze_a01_ruhela

Runs the ten-estimator DLRM (IPW, AIPW, g-computation, PSM 1:1, PSM subclassification, IRM-DML with ATE/ATTE/ATC, PSM-then-IRM-DML, ATC-AIPW, PLR, SuperLearner AIPW) on the canonical a01 formulation, T = alert-state complexity >= 2, Y = regional-volatility count, with an IRM-DML standard error comparison (pooled, by year, by person, two-way) and the Naive-arm sensitivity (any-flag treatment, binary outcome).

Usage

morie_otis_analyze_a01_ruhela_formulations(data = NULL, out_dir = NULL)

morie_otis_analyze_a01_dlrm(data = NULL, out_dir = NULL)

morie_otis_analyze_a01_mrm(data = NULL, out_dir = NULL)

Arguments

data: The OTIS a01 table. NULL uses morie_otis_load(): the table cached by morie_datasets_otis_a01(offline = FALSE), or else the included 5-row synthetic sample, with a warning that its numbers are not findings.
out_dir: Optional directory for the text and JSON rendering.

Value

morie_otis_analysis_result.

Examples

# \donttest{
a01 <- morie_synth_otis("a01", n = 300L, seed = 1L)
r <- morie_otis_analyze_a01_ruhela_formulations(a01)
r$tables[[1L]]$rows[[1L]]
#> [1] "1. IPW (Hajek)"                "ATE"                          
#> [3] "+2.0426"                       "0.1572"                       
#> [5] "[+1.735, +2.351]"              "1.27e-38"                     
#> [7] "calibration=none, Brier=0.173"
# }