enpls.fs: Ensemble Partial Least Squares for Feature Selection

In wind22zhu/enpls1.2: Ensemble Partial Least Squares Regression

Description

This function performs feature selection with ensemble partial least squares.

Usage

enpls.fs(x, y, maxcomp = NULL, MCtimes = 500L, method = c("mc",
  "bootstrap"), ratio = 0.8, parallel = 1L)

Arguments

x	predictor matrix
y	response vector
maxcomp	Maximum number of components included within the models, if not specified, default is the variable (column) numbers in x.
MCtimes	times of Monte-Carlo
method	"mc" or "bootstrap". Default is "mc".
ratio	sample ratio used when method = "mc"
parallel	Integer. Number of parallel processes to use. Default is 1, which means run serially.

Details

This function performs feature selection with ensemble partial least squares.

Value

A list containing two components:

• variable.importance - a vector of variable importance

• coefficient.matrix - original coefficient matrix

Author(s)

Min-feng Zhu <[e-mail]>, Nan Xiao <[e-mail]>

References

Dongsheng Cao, Yizeng Liang, Qingsong Xu, Yifeng Yun, and Hongdong Li. "Toward better QSAR/QSPR modeling: simultaneous outlier detection and variable selection using distribution of model features." Journal of computer-aided molecular design 25, no. 1 (2011): 67–80.

See Also

See enpls.od for outlier detection with ensemble PLS. See enpls.en for ensemble PLS regression. enpls.ad for applicability domain with ensemble PLS.

Examples

data(alkanes)
x = alkanes$x
y = alkanes$y

set.seed(42)
varimp = enpls.fs(x, y, MCtimes = 100)
print(varimp)
plot(varimp)

wind22zhu/enpls1.2 documentation built on May 4, 2019, 6:31 a.m.