PartitionTrainAndPredict: Partition Train And Predict

In citronmeliss/predictionpackr: Predictionpackr

Description

Partitions the study sample, fits the model and makes predictions with SuperLearner.

Usage

PartitionTrainAndPredict(study.sample,
  outcome.variable.name = "composite", model.names = c("ModelGAP",
  "ModelGerdin", "ModelKTS", "ModelRTS"), n.partitions = 2,
  save.to.results = TRUE, verbose = FALSE)

Arguments

study.sample	Data frame. The study.sample. No default.
outcome.variable.name	Character vector of length 1. The name of the outcome variable of interest. Defaults to "s30d".
n.partitions	Numeric vector of length 1. The number of partitions to create with PartitionSample. Accepted values are 2 or 3. If 2, a train and test set is created. If 3, train, validation, and test sets are created - the models is fitted on the training set, optimal breaks is gridsearched on the validation set, and the model is tested on the test set. Defaults to 2.
save.to.results	Logical. If TRUE SuperLearner predictions, outcome and tc in each partition is saved to the Results list. Defaults to TRUE.
verbose	Logical. If TRUE the modelling process is printed to console. Defaults to FALSE.
models.names	Character vector. The model names to stack in SuperLearner. Defaults to c("SL.gam", "SL.randomForest", "SL.nnet, SL.xgboost", "SL.svm")

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