train_gps --> estimate_gps

Author: Naeemkh

NAMESPACE

@@ -13,10 +13,10 @@ export(compute_rl_deriv_nn)
 export(compute_w_corr)
 export(estimate_cerf_gp)
 export(estimate_cerf_nngp)
+export(estimate_gps)
 export(generate_synthetic_data)
 export(get_logger)
 export(set_logger)
-export(train_gps)
 import(MASS)
 import(Rcpp)
 import(RcppArmadillo)

NEWS.md

@@ -1,6 +1,8 @@
 # GPCERF (developing)
 
 ## Changed
+- `estimate_gps` now returns the used exposure level, too. 
+- `train_gps` --> `estimate_gps` 
 - The nearest neighbor approach does not get `expand` as an input parameter (`n_neighbor` * `expand` --> `n_neighbor`).
 - The weighted covariate balance now is computed using the wCorr package.
 

R/compute_rl_deriv_gp.R

@@ -29,10 +29,10 @@
 #' \donttest{
 #' set.seed(847)
 #' data <- generate_synthetic_data(sample_size = 100)
-#' GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-#'                    w_all = data$treat,
-#'                    sl_lib = c("SL.xgboost"),
-#'                    dnorm_log = FALSE)
+#' GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+#'                       w_all = data$treat,
+#'                       sl_lib = c("SL.xgboost"),
+#'                       dnorm_log = FALSE)
 #'
 #' wi <- 8.6
 #'

R/compute_rl_deriv_nn.R

@@ -37,10 +37,10 @@
 #' \donttest{
 #' set.seed(325)
 #' data <- generate_synthetic_data(sample_size = 200)
-#' GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-#'                    w_all = data$treat,
-#'                    sl_lib = c("SL.xgboost"),
-#'                    dnorm_log = FALSE)
+#' GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+#'                       w_all = data$treat,
+#'                       sl_lib = c("SL.xgboost"),
+#'                       dnorm_log = FALSE)
 #'
 #' wi <- 12.2
 #'

R/estimate_cerf_gp.R

@@ -46,10 +46,10 @@
 #'
 #'
 #' # Estimate GPS function
-#' GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-#'                    w_all = data$treat,
-#'                    sl_lib = c("SL.xgboost"),
-#'                    dnorm_log = FALSE)
+#' GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+#'                       w_all = data$treat,
+#'                       sl_lib = c("SL.xgboost"),
+#'                       dnorm_log = FALSE)
 #'
 #' # exposure values
 #' w_all <- seq(0,10,1)

R/estimate_cerf_nngp.R

@@ -52,10 +52,10 @@
 #' set.seed(19)
 #' data <- generate_synthetic_data(sample_size = 120, gps_spec = 3)
 #' # Estimate GPS function
-#' GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-#'                    w_all = data$treat,
-#'                    sl_lib = c("SL.xgboost"),
-#'                    dnorm_log = FALSE)
+#' GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+#'                       w_all = data$treat,
+#'                       sl_lib = c("SL.xgboost"),
+#'                       dnorm_log = FALSE)
 #' # exposure values
 #' w.all <- seq(0,20,2)
 #' cerf_nngp_obj <- estimate_cerf_nngp(data,

R/train_GPS.R R/estimate_gps.R

@@ -1,9 +1,9 @@
 #' @title
-#' Train a model for generalized propensity score
+#' Estimate a model for generalized propensity score
 #'
 #' @description
-#' Estimates the conditional mean and sd of exposure level as a function of
-#' covariates.
+#' Estimates a model for generalized propensity score (GPS) using parametric
+#' approach.
 #'
 #' @param cov_mt A covariate matrix containing all covariates. Each row is a
 #' data sample and each column is a covariate.
@@ -23,13 +23,13 @@
 #' @examples
 #' \donttest{
 #' data <- generate_synthetic_data(sample_size = 200)
-#' GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-#'                   w_all = data$treat,
-#'                   sl_lib = c("SL.xgboost"),
-#'                   dnorm_log = FALSE)
+#' GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+#'                       w_all = data$treat,
+#'                       sl_lib = c("SL.xgboost"),
+#'                       dnorm_log = FALSE)
 #' }
 
-train_gps <- function(cov_mt, w_all, sl_lib, dnorm_log) {
+estimate_gps <- function(cov_mt, w_all, sl_lib, dnorm_log) {
 
   logger::log_info("Started estimating GPS values ... ")
   t_1 <- proc.time()

README.md

@@ -25,8 +25,10 @@ library("GPCERF")
 
   # Estimate GPS function
   # In the future, CausalGPS gps estimation will be used.
-  GPS_m <- train_GPS(cov_mt = as.matrix(sim_data[,-(1:2)]),
-                     w_all = as.matrix(sim_data$treat))
+  GPS_m <- estimate_gps(cov_mt = sim_data[,-(1:2)],
+                        w_all = sim_data$treat,
+                        sl_lib = c("SL.xgboost"),
+                        dnorm_log = FALSE)
 
   # exposure values
   w_all <- seq(0,20,0.1)

man/calc_ac.Rd

@@ -2,10 +2,10 @@ test_that("compute_deriv_nn works as expected!", {
 
   set.seed(765)
   data <- generate_synthetic_data(sample_size = 200)
-  GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-                     w_all = data$treat,
-                     sl_lib = c("SL.xgboost"),
-                     dnorm_log = FALSE)
+  GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+                        w_all = data$treat,
+                        sl_lib = c("SL.xgboost"),
+                        dnorm_log = FALSE)
 
   wi <- 4.8
 

man/compute_rl_deriv_gp.Rd

@@ -2,10 +2,10 @@ test_that("compute_deriv_weights_gp works as expected!", {
 
   set.seed(9615)
   data <- generate_synthetic_data(sample_size = 200)
-  GPS_m <- train_gps(cov_mt = data[, -(1:2)],
-                     w_all = data$treat,
-                     sl_lib = c("SL.xgboost"),
-                     dnorm_log = FALSE)
+  GPS_m <- estimate_gps(cov_mt = data[, -(1:2)],
+                        w_all = data$treat,
+                        sl_lib = c("SL.xgboost"),
+                        dnorm_log = FALSE)
 
   wi <- 4.2
   weights <- compute_deriv_weights_gp(w = wi,

man/compute_rl_deriv_nn.Rd

@@ -6,10 +6,10 @@ test_that("compute_m_sigma works as expected!", {
    w_all <- seq(0, 20, 0.1)
 
    #Estimate GPS function
-   GPS_m <- train_gps(cov_mt = data[, -(1:2)],
-                      w_all = data$treat,
-                      sl_lib = c("SL.xgboost"),
-                      dnorm_log = FALSE)
+   GPS_m <- estimate_gps(cov_mt = data[, -(1:2)],
+                         w_all = data$treat,
+                         sl_lib = c("SL.xgboost"),
+                         dnorm_log = FALSE)
 
    tune_res <- compute_m_sigma(hyperparam = c(0.09, 0.09, 10),
                                data = data,

man/estimate_cerf_gp.Rd

@@ -4,10 +4,10 @@ test_that("compute_posterior_m_nn works as expected.", {
   data <- generate_synthetic_data(sample_size = 200, gps_spec = 3)
 
   # Estimate GPS function
-  GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-                     w_all = data$treat,
-                     sl_lib = c("SL.xgboost"),
-                     dnorm_log = FALSE)
+  GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+                        w_all = data$treat,
+                        sl_lib = c("SL.xgboost"),
+                        dnorm_log = FALSE)
 
   # Hyperparameter
   hyperparam <- c(0.1, 0.2, 1)

man/estimate_cerf_nngp.Rd

@@ -4,10 +4,10 @@ test_that("compute_posterior_sd_nn works as expected.", {
   data <- generate_synthetic_data(sample_size = 200, gps_spec = 3)
 
   # Estimate GPS function
-  GPS_m <- train_gps(cov_mt = data[, -(1:2)],
-                     w_all = data$treat,
-                     sl_lib = c("SL.xgboost"),
-                     dnorm_log = FALSE)
+  GPS_m <- estimate_gps(cov_mt = data[, -(1:2)],
+                        w_all = data$treat,
+                        sl_lib = c("SL.xgboost"),
+                        dnorm_log = FALSE)
 
   # Hyperparameter
   hyperparam <- c(0.1, 0.2, 1)

man/train_gps.Rd man/estimate_gps.Rd

@@ -2,10 +2,10 @@ test_that("compute_rl_deriv_gp works as expected!", {
 
   set.seed(127)
   data <- generate_synthetic_data(sample_size = 200)
-  GPS_m <- train_gps(cov_mt = data[, -(1:2)],
-                     w_all = data$treat,
-                     sl_lib = c("SL.xgboost"),
-                     dnorm_log = FALSE)
+  GPS_m <- estimate_gps(cov_mt = data[, -(1:2)],
+                        w_all = data$treat,
+                        sl_lib = c("SL.xgboost"),
+                        dnorm_log = FALSE)
 
   wi <- 8.6
 

tests/testthat/test-calc_ac.R

@@ -2,10 +2,10 @@ test_that("compute_rl_deriv_nn works as expected!", {
 
   set.seed(325)
   data <- generate_synthetic_data(sample_size = 200)
-  GPS_m <- train_gps(cov_mt = data[, -(1:2)],
-                     w_all = data$treat,
-                     sl_lib = c("SL.xgboost"),
-                     dnorm_log = FALSE)
+  GPS_m <- estimate_gps(cov_mt = data[, -(1:2)],
+                        w_all = data$treat,
+                        sl_lib = c("SL.xgboost"),
+                        dnorm_log = FALSE)
 
   wi <- 12.2
 

tests/testthat/test-compute_deriv_nn.R

@@ -11,10 +11,10 @@ test_that("compute_sd_gp works as expected.", {
    kernel_fn <- function(x) exp(-x^2)
 
    # Estimate GPS function
-   GPS_m <- train_gps(cov_mt = data[,-(1:2)],
-                      w_all = data$treat,
-                      sl_lib = c("SL.xgboost"),
-                      dnorm_log = FALSE)
+   GPS_m <- estimate_gps(cov_mt = data[,-(1:2)],
+                         w_all = data$treat,
+                         sl_lib = c("SL.xgboost"),
+                         dnorm_log = FALSE)
 
    GPS <- GPS_m$gps$GPS