Renaming argument prediction_zero to phi0 (#413)

NorskRegnesentral · Oct 22, 2024 · e2caa4a · e2caa4a
1 parent 03a4206
commit e2caa4a
Show file tree

Hide file tree

Showing 203 changed files with 1,095 additions and 1,094 deletions.
diff --git a/NAMESPACE b/NAMESPACE
@@ -54,6 +54,7 @@ S3method(setup_approach,timeseries)
 S3method(setup_approach,vaeac)
 export(additional_regression_setup)
 export(aicc_full_single_cpp)
+export(append_vS_list)
 export(check_convergence)
 export(cli_compute_vS)
 export(cli_iter)

diff --git a/R/approach_vaeac.R b/R/approach_vaeac.R
@@ -2531,7 +2531,7 @@ Last epoch:             %d. \tVLB = %.3f \tIWAE = %.3f \tIWAE_running = %.3f\n",
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = approach,
-#'   prediction_zero = p0,
+#'   phi0 = p0,
 #'   n_MC_samples = 1, # As we are only interested in the training of the vaeac
 #'   vaeac.epochs = 10, # Should be higher in applications.
 #'   vaeac.n_vaeacs_initialize = 1,
@@ -2545,7 +2545,7 @@ Last epoch:             %d. \tVLB = %.3f \tIWAE = %.3f \tIWAE_running = %.3f\n",
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = approach,
-#'   prediction_zero = p0,
+#'   phi0 = p0,
 #'   n_MC_samples = 1, # As we are only interested in the training of the vaeac
 #'   vaeac.epochs = 10, # Should be higher in applications.
 #'   vaeac.width = 16,
@@ -2766,7 +2766,7 @@ vaeac_plot_eval_crit <- function(explanation_list,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "vaeac",
-#'   prediction_zero = mean(y_train),
+#'   phi0 = mean(y_train),
 #'   n_MC_samples = 1,
 #'   vaeac.epochs = 10,
 #'   vaeac.n_vaeacs_initialize = 1

diff --git a/R/cli.R b/R/cli.R
@@ -80,7 +80,6 @@ cli_startup <- function(internal, model_class, verbose) {
 #' @export
 #' @keywords internal
 cli_compute_vS <- function(internal) {
-
   verbose <- internal$parameters$verbose
   approach <- internal$parameters$approach
 

diff --git a/R/compute_estimates.R b/R/compute_estimates.R
@@ -76,11 +76,11 @@ compute_estimates <- function(internal, vS_list) {
 #' @keywords internal
 postprocess_vS_list <- function(vS_list, internal) {
   keep_samp_for_vS <- internal$parameters$output_args$keep_samp_for_vS
-  prediction_zero <- internal$parameters$prediction_zero
+  phi0 <- internal$parameters$phi0
   n_explain <- internal$parameters$n_explain
 
   # Appending the zero-prediction to the list
-  dt_vS0 <- as.data.table(rbind(c(1, rep(prediction_zero, n_explain))))
+  dt_vS0 <- as.data.table(rbind(c(1, rep(phi0, n_explain))))
 
   # Extracting/merging the data tables from the batch running
   # TODO: Need a memory and speed optimized way to transform the output form dt_vS_list to two different lists,

diff --git a/R/compute_vS.R b/R/compute_vS.R
@@ -230,7 +230,6 @@ compute_MCint <- function(dt, pred_cols = "p_hat") {
 #' @export
 #' @keywords internal
 append_vS_list <- function(vS_list, internal) {
-
   iter <- length(internal$iter_list)
 
   # Adds v_S output above to any vS_list already computed
@@ -243,17 +242,17 @@ append_vS_list <- function(vS_list, internal) {
 
     # Creates a mapper from the last id_coalition to the new id_coalition numbering
     id_coalitions_mapper <- merge(prev_coalition_map,
-                                  current_coalition_map,
-                                  by = "coalitions_str",
-                                  suffixes = c("", "_new")
+      current_coalition_map,
+      by = "coalitions_str",
+      suffixes = c("", "_new")
     )
     prev_vS_list_new <- list()
 
     # Applies the mapper to update the prev_vS_list ot the new id_coalition numbering
     for (k in seq_along(prev_vS_list)) {
       prev_vS_list_new[[k]] <- merge(prev_vS_list[[k]],
-                                     id_coalitions_mapper[, .(id_coalition, id_coalition_new)],
-                                     by = "id_coalition"
+        id_coalitions_mapper[, .(id_coalition, id_coalition_new)],
+        by = "id_coalition"
       )
       prev_vS_list_new[[k]][, id_coalition := id_coalition_new]
       prev_vS_list_new[[k]][, id_coalition_new := NULL]
@@ -263,5 +262,4 @@ append_vS_list <- function(vS_list, internal) {
     vS_list <- c(prev_vS_list_new, vS_list)
   }
   return(vS_list)
-
 }
diff --git a/R/explain.R b/R/explain.R
@@ -21,7 +21,7 @@
 #' `"categorical"`, `"timeseries"`, `"independence"`, `"regression_separate"`, or `"regression_surrogate"`.
 #' The two regression approaches can not be combined with any other approach. See details for more information.
 #'
-#' @param prediction_zero Numeric.
+#' @param phi0 Numeric.
 #' The prediction value for unseen data, i.e. an estimate of the expected prediction without conditioning on any
 #' features.
 #' Typically we set this value equal to the mean of the response variable in our training data, but other choices
@@ -211,7 +211,7 @@
 #' \describe{
 #'   \item{shapley_values_est}{data.table with the estimated Shapley values with explained observation in the rows and
 #'   features along the columns.
-#'   The column `none` is the prediction not devoted to any of the features (given by the argument `prediction_zero`)}
+#'   The column `none` is the prediction not devoted to any of the features (given by the argument `phi0`)}
 #'   \item{shapley_values_sd}{data.table with the standard deviation of the Shapley values reflecting the uncertainty.
 #'   Note that this only reflects the coalition sampling part of the kernelSHAP procedure, and is therefore by
 #'   definition 0 when all coalitions is used.
@@ -269,7 +269,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "empirical",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -279,7 +279,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "gaussian",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -289,7 +289,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "copula",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -299,7 +299,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "ctree",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -310,7 +310,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = approach,
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -332,7 +332,7 @@
 #'   x_train = x_train,
 #'   group = group_list,
 #'   approach = "empirical",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #' print(explain_groups$shapley_values_est)
@@ -347,7 +347,7 @@
 #'   model = model,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   approach = "regression_separate",
 #'   regression.model = parsnip::linear_reg()
 #' )
@@ -356,7 +356,7 @@
 #'   model = model,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   approach = "regression_surrogate",
 #'   regression.model = parsnip::linear_reg()
 #' )
@@ -370,7 +370,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "gaussian",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2,
 #'   iterative = TRUE,
 #'   iterative_args = list(initial_n_coalitions = 10)
@@ -395,7 +395,7 @@ explain <- function(model,
                     x_explain,
                     x_train,
                     approach,
-                    prediction_zero,
+                    phi0,
                     iterative = NULL,
                     max_n_coalitions = NULL,
                     group = NULL,
@@ -433,7 +433,7 @@ explain <- function(model,
     x_explain = x_explain,
     approach = approach,
     paired_shap_sampling = paired_shap_sampling,
-    prediction_zero = prediction_zero,
+    phi0 = phi0,
     max_n_coalitions = max_n_coalitions,
     group = group,
     n_MC_samples = n_MC_samples,

diff --git a/R/explain_forecast.R b/R/explain_forecast.R
@@ -79,7 +79,7 @@
 #'   explain_y_lags = 2,
 #'   horizon = 3,
 #'   approach = "empirical",
-#'   prediction_zero = p0_ar,
+#'   phi0 = p0_ar,
 #'   group_lags = FALSE
 #' )
 #'
@@ -93,7 +93,7 @@ explain_forecast <- function(model,
                              explain_xreg_lags = explain_y_lags,
                              horizon,
                              approach,
-                             prediction_zero,
+                             phi0,
                              max_n_coalitions = NULL,
                              iterative = NULL,
                              iterative_args = list(),
@@ -125,7 +125,7 @@ explain_forecast <- function(model,
   # Checks data/model compatability
   internal <- setup(
     approach = approach,
-    prediction_zero = prediction_zero,
+    phi0 = phi0,
     output_size = horizon,
     max_n_coalitions = max_n_coalitions,
     n_MC_samples = n_MC_samples,

diff --git a/R/plot.R b/R/plot.R
@@ -101,7 +101,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "empirical",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -151,7 +151,7 @@
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "ctree",
-#'   prediction_zero = p,
+#'   phi0 = p,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -299,7 +299,7 @@ plot.shapr <- function(x,
     # compute start and end values for waterfall rectangles
     data.table::setorder(dt_plot, rank_waterfall)
     dt_plot[, end := cumsum(phi), by = id]
-    expected <- x$internal$parameters$prediction_zero
+    expected <- x$internal$parameters$phi0
     dt_plot[, start := c(expected, head(end, -1)), by = id]
     dt_plot[, phi_significant := format(phi, digits = digits), by = id]
 
@@ -895,15 +895,15 @@ make_waterfall_plot <- function(dt_plot,
 #' )
 #'
 #' # Specifying the phi_0, i.e. the expected prediction without any features
-#' prediction_zero <- mean(y_train)
+#' phi0 <- mean(y_train)
 #'
 #' # Independence approach
 #' explanation_independence <- explain(
 #'   model = model,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "independence",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -913,7 +913,7 @@ make_waterfall_plot <- function(dt_plot,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "gaussian",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e1
 #' )
 #'
@@ -923,7 +923,7 @@ make_waterfall_plot <- function(dt_plot,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "gaussian",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -933,7 +933,7 @@ make_waterfall_plot <- function(dt_plot,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "ctree",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -943,7 +943,7 @@ make_waterfall_plot <- function(dt_plot,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = c("gaussian", "independence", "ctree"),
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -1452,15 +1452,15 @@ make_MSEv_coalition_plots <- function(MSEv_coalition_dt,
 #' )
 #'
 #' # Specifying the phi_0, i.e. the expected prediction without any features
-#' prediction_zero <- mean(y_train)
+#' phi0 <- mean(y_train)
 #'
 #' # Independence approach
 #' explanation_independence <- explain(
 #'   model = model,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "independence",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -1470,7 +1470,7 @@ make_MSEv_coalition_plots <- function(MSEv_coalition_dt,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "empirical",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -1480,7 +1480,7 @@ make_MSEv_coalition_plots <- function(MSEv_coalition_dt,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "gaussian",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e1
 #' )
 #'
@@ -1490,7 +1490,7 @@ make_MSEv_coalition_plots <- function(MSEv_coalition_dt,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = "gaussian",
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'
@@ -1500,7 +1500,7 @@ make_MSEv_coalition_plots <- function(MSEv_coalition_dt,
 #'   x_explain = x_explain,
 #'   x_train = x_train,
 #'   approach = c("gaussian", "ctree", "empirical"),
-#'   prediction_zero = prediction_zero,
+#'   phi0 = phi0,
 #'   n_MC_samples = 1e2
 #' )
 #'