create.ROCR.plots.v2: Create ROCR plots

In citronmeliss/predictionpackr: Predictionpackr

Description

Plots reciever operating charecteristiqs curves or precision/recall curves.

Usage

create.ROCR.plots.v2(study_sample, outcome_name, device = "eps",
  split_var = "train", train_test = TRUE, ROC_or_precrec = "ROC",
  models = c("pred_con_train", "pred_cat_train", "pred_con_test",
  "pred_cat_test", "tc"),
  pretty_names = c("SuperLearner continuous prediction",
  "SuperLearner priority levels", "SuperLearner continuous prediction",
  "SuperLearner priority levels", "Clinicians priority levels"),
  subscript = FALSE, models_to_invert = NULL)

Arguments

study_sample	Study sample list. No default.
device	Character vector with the name of the image device to use. Passed to rocr.plot (in turn, passed to save.plot). Defaults to "eps".
split_var	The variable used to split plots. As string. Defaults to "train".
train_test	Logical. Is the dataset splitted in train and test set? Defaults to TRUE.
ROC_or_precrec	String. To perform ROC or precision/recall analysis. Accepted values are "ROC" or "prec_rec". No default.
models	Model names as character vector. Defaults to c("pred_con_train", "pred_cat_train", "pred_con_test", "pred_cat_test", "tc")
pretty_names	Names to be used in plots. As character vector. Defaults to c("SuperLearner continuous prediction", "SuperLearner priority levels", "SuperLearner continuous prediction", "SuperLearner priority levels", "Clinicians priority levels")
subscript	Logical. If TRUE, underscores in pretty names in converted to expression. Passed to rocr.plots. Defaults to FALSE.
models_to_invert	Character vector. Names of models to invert. Defaults to NULL.

citronmeliss/predictionpackr documentation built on Feb. 10, 2020, 12:19 a.m.