preprosimrun: Run simulation

In preprosim: Lightweight Data Quality Simulation for Classification

Description

Run simulation

Usage

preprosimrun(data, param = newparam(data, "default"), seed = 1,
  caretmodel = "gbm", holdoutrounds = 10, cores = 1, verbose = TRUE,
  fitmodels = TRUE)

Arguments

data	(data frame) one factor columns for class labels, other columns numeric, no missing values
param	(preprosimparameter object) simulation parameters, defaults to parameters set automatically for data.
seed	(integer) seed to be used for reproducible results, defaults to 1
caretmodel	(character) a model from package Caret, defaults to gbm (gbm must be installed before preprosimrun)
holdoutrounds	(integer) number of holdout rounds, defaults to 10
cores	(integer) number of cores used in parallel processing, defaults to 1
verbose	(boolean) progress information outputted, defaults to TRUE
fitmodels	(boolean) whether classification models are fitted, defaults to TRUE (FALSE: get only the contaminated datasets)

Details

caretmodel must be able to deal with missing values and have in-build variable importance such as rpart and gbm. Note: caret message will be outputted regardless of verbose.

Value

preprosimanalysis class object

Examples

res <- preprosimrun(iris, param=newparam(iris, "custom", x="misval", z="noise"), fitmodels=FALSE)

preprosim documentation built on May 1, 2019, 6:27 p.m.