Perform CARS-PLS analysis.

This function performs Competitive Adaptive Reweighted Sampling (CARS) for partial least squares (PLS) analysis.

Usage

get.cars.pls(
  X,
  y,
  nLV = 2,
  fold = 10,
  scale.pretreat = 1,
  iteration = 50,
  PartitionType = "interleaved"
)

Arguments

X

The sample matrix, where samples are in rows and variables are in columns.

y

The response variable.

nLV

The number of latent variables in PLS. Default is 2.

fold

The number of segments for cross-validation. Default is 10.

scale.pretreat

Whether to scale the variables. 1 for scaling, 0 for no scaling (only centered). Default is 1.

iteration

The number of Monte Carlo samplings in CARS. Default is 50.

PartitionType

The partition type for cross-validation: "random", "consecutive", or "interleaved". Default is "interleaved".

Value

A list containing the results of the CARSPLS analysis.

Author

Ported by Carlos Camino; orignal code oin matlab by Yizeng Liang, and Hongdong Li

Examples

if (FALSE) { # \dontrun{
# X: sample matrix (rows = samples, columns = variables); y: response vector
get.cars.pls(X, y)
get.cars.pls(X, y, nLV = 3, fold = 5, scale.pretreat = 0)
} # }