get_loadings: Get loadings

In jonathanth/mixOmicsCaret: mixOmicsCaret

Get loadings

Description

Returns the loadings of a mixOmics spls train object. Either from the final model fitted on the full training data, or the median and sd across CV folds from the best performing tuning parameters. If there are multiple CV repeats, these are separated or can be chosen.

Usage

get_loadings(trainobj, what = c("finalModel", "CV"), xykeep = c("x", "y",
  "both"), rep = NA, remove_empty = TRUE, ncomp = NA, keepX = NA,
  keepY = NA)

Arguments

trainobj	The train object from caret, fit using mixOmics spls.
what	Which model to return loadings from. Either "finalModel" or "CV".
xykeep	Return loadings from x or y? Either "x", "y", or "both".
rep	If what = "CV", and trainobj was fit using repeated cross-validation, choose a repeat (e.g. "Rep1") or leave as NA for all repeats (default).
remove_empty	Remove loadings with a value of zero from output.
ncomp	Manually select CV predictions with this parameter.
keepX	Manually select CV predictions with this parameter.
keepY	Manually select CV predictions with this parameter.

Value

A data.frame with the variables

• var - the variable

• comp - the component

• loading - the loading value (median if across CV folds / reps)

• sd - the standard deviation of the loading (if across CV folds / reps, NA otherwise)

• xy - is it a loading on x or y?

Examples

library(caret)
x <- data.frame(matrix(rnorm(1000),nrow = 100))
y <- rnorm(100)
PLS <- train(x = x, y = y, method = get_mixOmics_spls())
get_loadings(PLS)

get_loadings(PLS, "CV") %>%
  ggplot(aes(var, loading, ymin = loading - sd, ymax = loading + sd)) +
    facet_wrap(~ comp, scales = "free_x") +
    geom_errorbar() +
    geom_bar(stat = "identity")

jonathanth/mixOmicsCaret documentation built on Feb. 25, 2023, 5:41 a.m.