sampling_bootstrap: Sampling procedures used for testing capacity algorithm

In SLEMI: Statistical Learning Based Estimation of Mutual Information

Sampling procedures used for testing capacity algorithm

Description

Internal, auxiliary functions

Usage

sampling_bootstrap(data, prob = 1, dataDiv)

sampling_shuffle(data, side_variables)

sampling_partition(data, dataDiv, partition_trainfrac)

Arguments

data	is a data.frame to be resampled
prob	is numeric for the portion of data that should be sampled from the whole dataset (only in sampling_bootstrap)
dataDiv	a character indicating column of data, with respect to which, data should be split before bootstrap
side_variables	is a vector of characters indicating columns of data the will be reshuffled (only in sampling_shuffle)
partition_trainfrac	is a numeric for the portion of data that will be used as a training and testing datasets

Details

These function allow to re-sample, bootstrap and divide initial dataset

Value

Function sampling_bootstrap returns a data.frame with the same structure as initial data object, but with prob proportion of observations for each dataDiv level. Function sampling_shuffle returns a data.frame with the same structure as initial data object with shuffled values of columns given in side_variables argument. Function sampling_partition returns a list of two data.frame objects - train and test that has the same structure as initial data argument with partition_trainfrac and 1-partition_trainfrac observations, respectively.

Examples

data=data_example1
dataBootstrap = SLEMI:::sampling_bootstrap(data=data,prob=0.8,data$signal)
dataShuffle = SLEMI:::sampling_shuffle(data=data,"sideVar")
dataTrainTest = SLEMI:::sampling_partition(data=data,dataDiv=data$signal,partition_trainfrac=0.6)

SLEMI documentation built on Nov. 20, 2023, 1:06 a.m.