bp_VIP_analysis: Bootstrap and permutation over PLS-VIP

In sipss/NIHSlcms: Signal processing of LC-MS metabolomics

Description

Bootstrap and permutation over PLS-VIP performed on lcms_dataset peak tables.

Usage

bp_VIP_analysis(
  dataset,
  train_index,
  y_column,
  ncomp,
  nbootstrap = 300,
  multilevel = NULL
)

Arguments

dataset	An lcms_dataset object
train_index	set of index used to generate the bootstrap datasets
y_column	A list with the clases of the samples
ncomp	number of components used in the plsda models
nbootstrap	number of bootstrap dataset

Details

Use of the bootstrap and permutation methods for a more robust variable importance in the projection metric for partial least squares regression

Value

A list with the following elements:

• important_vips: A list with the important vips selected

• relevant_vips: List of vips with some relevance

• pls_vip: Pls-VIPs of every bootstrap

• pls_vip_perm: Pls-VIPs of every bootstrap with permuted variables

• pls_vip_means: Pls-VIPs normaliced differences means

• pls_vip_score_diff: Differences of pls_vip and pls_vip_perm

• pls_models: pls models of the diferent bootstraps

• pls_perm_models: pls permuted models of the diferent bootstraps

• classif_rate: classification rate of the bootstrap models

• general_model: pls model trained with all train data

• general_CR: classification rate of the general_model

• vips_model: pls model trained with vips selection over all train data

• vips_CR: classification rate of the vips_model

• error: error spected in a t distribution

• lower_bound: lower bound of the confidence interval

• upper_bound: upper bound of the confidence interval

sipss/NIHSlcms documentation built on May 13, 2021, 6:19 p.m.