newparam: Create new simulation control parameter object

In mvattulainen/preprosim: Lightweight Data Quality Simulation for Classification

Description

Preprosim parameter objects contain eight contaminations: noise, lowvar, misval, irfeature, classswap, classimbalance, volumedecrease and outlier. Each contamination has three sub parameters: cols as columns the contamination is applied to, param as the parameter of the contamination itself (i.e. intensity of contamination) and order as order in which the parameter is applied to the data.

Usage

newparam(dataframe, type = "default", x, z)

Arguments

dataframe	(data frame) original data to be used in simulations
type	(character) creation type: empty, default or custom, defaults to "default"
x	(character) primary contamination of interest such as "misval"
z	(character) secondary contamination of interest such as "noise"

Details

For argument type: empty creates a preprosimparameter object with empty params (but not empty cols or order). default creates 6561 combinations with all params 0, 0.1, 0.2. custom creates params seq(0, 0.9, by 0.1) for primary (x) and 0., 0.1, 0.2 for secondary (z). The implicit y (not an argument) refers to classification accuracy.

Value

preprosimparameter class object

Examples

pa <- newparam(iris)
pa1 <- newparam(iris, "empty")
pa2 <- newparam(iris, "custom", "misval", "noise")

mvattulainen/preprosim documentation built on May 23, 2019, 10:54 a.m.