ViP: Variable importance in projection (VIP)

In fastPLS: Fast Partial Least Squares for High-Dimensional Data

Variable importance in projection (VIP)

Description

Computes VIP trajectories from fitted direct SIMPLS-family components. The standard component-wise decomposition is not used for PLS-SVD, OPLS, or nonlinear kernel PLS because their stored latent weights have different mathematical meanings. Linear-kernel PLS uses the same direct SIMPLS-family path and is supported.

Usage

ViP(model)

Arguments

model

Fitted fastPLS model.

Value

Numeric matrix (single response) or list of matrices (multi-response).

Examples

X <- as.matrix(mtcars[, c("disp", "hp", "wt", "qsec")])
y <- mtcars$mpg
fit <- pls(X, y,
    ncomp = 1, method = "simpls", backend = "cpu",
    fit = TRUE, return_variance = FALSE
)
ViP(fit)

fastPLS documentation built on Sept. 29, 2026, 1:06 a.m.