Merge pull request #243 from NSAPH-Software/bug_fix_4

Bug fix 4

.gitignore

@@ -2,6 +2,8 @@
 .Rproj.user/*
 .Rhistory
 .RData
+*.RData
+*.rds
 .Ruserdata
 inst/doc
 tests/testthat/_snaps/*
@@ -17,4 +19,4 @@ src-i386
 functional_tests/supplements/*
 *zip_data.RData
 *log*
-*.rds
+*a.rds

R/estimate_erf.R

@@ -207,8 +207,6 @@ estimate_erf <- function(.data,
     result_data_prediction <- data.frame(w_vals = w_vals,
                                          y_pred = y_pred)
 
-    print(erf_np)
-
   } else {
     stop("The code should never get here. Double check.")
   }

R/estimate_gps.R

@@ -18,7 +18,8 @@
 #'
 #' @return
 #' The function returns a S3 object. Including the following:
-#'   - `.data `: `id`, `w`, `gps`, `e_gps_pred`, `e_gps_std_pred`, `w_resid`
+#'   - `.data `: `id`, `exposure_var`, `gps`, `e_gps_pred`, `e_gps_std_pred`,
+#'   `w_resid`
 #'   - `params`: Including the following fields:
 #'     - gps_mx (min and max of gps)
 #'     - w_mx (min and max of w).
@@ -113,7 +114,10 @@ estimate_gps <- function(.data,
   gps_mx <- compute_min_max(gps)
 
   # create new data.frame for output
-  dataset <- data.frame(id = .data$id, w = response_data, gps = gps)
+  dataset <- stats::setNames(data.frame(.data$id,
+                                        response_data,
+                                        gps),
+                      c("id", response_var, "gps"))
   dataset$e_gps_pred <- e_gps_pred
   if (length(e_gps_std_pred) == 1){
     e_gps_std_pred <- rep(e_gps_std_pred, nrow(dataset))

R/matching_fn.R

@@ -75,7 +75,7 @@ matching_fn <- function(w,
   std_w <- (w - w_min) / (w_max - w_min)
   std_p_w <- (p_w - gps_min) / (gps_max - gps_min)
 
-  dataset_subset <- dataset[abs(dataset[["w"]] - w) <= (delta_n / 2), ]
+  dataset_subset <- dataset[abs(dataset[[exposure_col_name]] - w) <= (delta_n / 2), ]
 
   if (nrow(dataset_subset) < 1){
     logger:: log_warn(paste("There is no data to match with ", w, "in ",

functional_tests/paper_1_estimating_gps_normal.R

@@ -0,0 +1,70 @@
+##      author:  Naeem Khoshnevis
+##      created: September 2023
+##      purpose: Reproducing examples in the paper.
+
+
+# Load libraries
+library(ggplot2)
+library(CausalGPS)
+library(data.table)
+
+# Load data --------------------------------------------------------------------
+data_file <- "zip_data.RData"
+if (!file.exists(data_file)) {
+  stop(paste0("Download the study data file from the following link:\n",
+              "https://drive.google.com/file/d/",
+              "1QFdbVU8Qir1gWf96c5h_ZhT-aPjhHpqn/view?usp=share_link"))
+} else {
+  load(data_file)
+}
+
+data.table::setDF(zip_data)
+data <- zip_data
+
+# Add id to the data
+data$id <- 1:nrow(data)
+data$w <- data$pm25
+data$pm25 <- NULL
+
+save(data, file = "study_data.RData")
+
+# Estimate GPS -----------------------------------------------------------------
+
+## Super learner wrapper
+m_xgboost <- function(nthread = 6,
+                      ntrees = 50,
+                      shrinkage = 0.3,
+                      max_depth = 6,
+                      minobspernode = 1,
+                      verbose = 1,
+                      ...) {SuperLearner::SL.xgboost(
+                        nthread = nthread,
+                        ntrees = ntrees,
+                        shrinkage=shrinkage,
+                        max_depth=max_depth,
+                        mibobspernode=minobspernode,
+                        verbose=verbose,
+                        ...)}
+
+exposure <- "w"
+confounders   <- c("mean_bmi", "smoke_rate",
+                   "hispanic", "pct_blk", "medhouseholdincome",
+                   "medianhousevalue", "poverty", "popdensity",
+                   "pct_owner_occ", "summer_tmmx", "winter_tmmx",
+                   "summer_rmax", "winter_rmax", "year")
+
+formula_str <- paste(exposure, " ~ ", paste(confounders, collapse = " + "))
+
+
+data_with_gps_normal <- estimate_gps(.data = data,
+                                     .formula = as.formula(formula_str),
+                                     gps_density = "normal",
+                                     sl_lib = c("m_xgboost")
+)
+
+
+pdf("figure_paper_1_estimating_gps_normal.pdf")
+plot(data_with_gps_normal)
+dev.off()
+
+save(data_with_gps_normal, file = "data_with_gps_normal.RData")

functional_tests/paper_2_estimating_gps_kernel.R

@@ -0,0 +1,66 @@
+##      author:  Naeem Khoshnevis
+##      created: March 2024
+##      purpose: Reproducing examples in the paper.
+
+
+# Load libraries
+library(ggplot2)
+library(CausalGPS)
+library(data.table)
+
+# Load data --------------------------------------------------------------------
+data_file <- "zip_data.RData"
+if (!file.exists(data_file)) {
+  stop(paste0("Download the study data file from the following link:\n",
+              "https://drive.google.com/file/d/",
+              "1QFdbVU8Qir1gWf96c5h_ZhT-aPjhHpqn/view?usp=share_link"))
+} else {
+  load(data_file)
+}
+
+data.table::setDF(zip_data)
+data <- zip_data
+
+# Add id to the data
+data$id <- 1:nrow(data)
+
+# Estimate GPS -----------------------------------------------------------------
+
+## Super learner wrapper
+m_xgboost <- function(nthread = 6,
+                      ntrees = 50,
+                      shrinkage = 0.3,
+                      max_depth = 6,
+                      minobspernode = 1,
+                      verbose = 1,
+                      ...) {SuperLearner::SL.xgboost(
+                        nthread = nthread,
+                        ntrees = ntrees,
+                        shrinkage=shrinkage,
+                        max_depth=max_depth,
+                        mibobspernode=minobspernode,
+                        verbose=verbose,
+                        ...)}
+
+exposure <- "pm25"
+confounders   <- c("mean_bmi", "smoke_rate",
+                   "hispanic", "pct_blk", "medhouseholdincome",
+                   "medianhousevalue", "poverty", "popdensity",
+                   "pct_owner_occ", "summer_tmmx", "winter_tmmx",
+                   "summer_rmax", "winter_rmax", "year")
+
+formula_str <- paste(exposure, " ~ ", paste(confounders, collapse = " + "))
+
+
+data_with_gps_kernel <- estimate_gps(.data = data,
+                                     .formula = as.formula(formula_str),
+                                     gps_density = "kernel",
+                                     sl_lib = c("m_xgboost")
+                                    )
+
+
+pdf("figure_paper_2_estimating_gps_kernel.pdf")
+plot(data_with_gps_kernel)
+dev.off()
+
+save(data_with_gps_kernel, file = "data_with_gps_kernel.RData")

functional_tests/paper_3_compute_weight_counter_weighting.R

@@ -0,0 +1,26 @@
+##      author:  Naeem Khoshnevis
+##      created: March 2024
+##      purpose: Reproducing examples in the paper.
+
+
+# Load libraries
+library(ggplot2)
+library(CausalGPS)
+
+
+# Load gps object
+load("data_with_gps_normal.RData")
+
+
+cw_weighting_object <- compute_counter_weight(gps_obj = data_with_gps_normal,
+                                              ci_appr = "weighting",
+                                              bin_seq = NULL,
+                                              nthread = 6,
+                                              delta_n = 0.1,
+                                              dist_measure = "l1",
+                                              scale = 0.5)
+
+
+save(cw_weighting_object, file = "cw_weighting_object.RData")
+
+

functional_tests/paper_4_compute_weight_counter_matching.R

@@ -0,0 +1,24 @@
+##      author:  Naeem Khoshnevis
+##      created: March 2024
+##      purpose: Reproducing examples in the paper.
+
+
+# Load libraries
+library(ggplot2)
+library(CausalGPS)
+
+
+# Load gps object
+load("data_with_gps_normal.RData")
+
+
+cw_matching_object <- compute_counter_weight(gps_obj = data_with_gps_normal,
+                                             ci_appr = "matching",
+                                             bin_seq = NULL,
+                                             nthread = 6,
+                                             delta_n = 0.1,
+                                             dist_measure = "l1",
+                                             scale = 0.5)
+
+
+save(cw_matching_object, file = "cw_matching_object.RData")

functional_tests/paper_5_compute_pseudo_pop_weighting.R

@@ -0,0 +1,36 @@
+##      author:  Naeem Khoshnevis
+##      created: March 2024
+##      purpose: Reproducing examples in the paper.
+
+
+# Load libraries
+library(ggplot2)
+library(CausalGPS)
+
+
+# Load cw object and data
+load("cw_weighting_object.RData")
+load("study_data.RData")
+
+confounders   <- c("mean_bmi", "smoke_rate",
+                   "hispanic", "pct_blk", "medhouseholdincome",
+                   "medianhousevalue", "poverty", "popdensity",
+                   "pct_owner_occ", "summer_tmmx", "winter_tmmx",
+                   "summer_rmax", "winter_rmax", "year")
+
+
+pseudo_pop_weighting_object <- generate_pseudo_pop(
+                                           .data = data,
+                                           cw_obj = cw_weighting_object,
+                                           covariate_col_names = confounders,
+                                           covar_bl_trs = 0.1,
+                                           covar_bl_trs_type = "maximal",
+                                           covar_bl_method = "absolute")
+
+
+save(pseudo_pop_weighting_object, file = "pseudo_pop_weighting_object.RData")
+
+
+pdf("figure_paper_5_pseudo_pop_weighting_object.pdf")
+plot(pseudo_pop_weighting_object)
+dev.off()

functional_tests/paper_6_compute_pseudo_pop_matching.R

@@ -0,0 +1,35 @@
+##      author:  Naeem Khoshnevis
+##      created: March 2024
+##      purpose: Reproducing examples in the paper.
+
+
+# Load libraries
+library(ggplot2)
+library(CausalGPS)
+
+
+# Load cw object and data
+load("cw_matching_object.RData")
+load("study_data.RData")
+
+confounders   <- c("mean_bmi", "smoke_rate",
+                   "hispanic", "pct_blk", "medhouseholdincome",
+                   "medianhousevalue", "poverty", "popdensity",
+                   "pct_owner_occ", "summer_tmmx", "winter_tmmx",
+                   "summer_rmax", "winter_rmax", "year")
+
+
+pseudo_pop_weighting_object <- generate_pseudo_pop(
+                                           .data = data,
+                                           cw_obj = cw_matching_object,
+                                           covariate_col_names = confounders,
+                                           covar_bl_trs = 0.1,
+                                           covar_bl_trs_type = "maximal",
+                                           covar_bl_method = "absolute")
+
+
+save(pseudo_pop_matching_object, file = "pseudo_pop_matching_object.RData")
+
+pdf("figure_paper_6_pseudo_pop_matching_object.pdf")
+plot(pseudo_pop_matching_object)
+dev.off()