| ispls {iSFun} | R Documentation |
Integrative sparse partial least squares
Description
This function provides a penalty-based integrative sparse partial least squares method to handle the multiple datasets with high dimensions generated under similar protocols, which consists of two built-in penalty items for selecting the important variables for users to choose, and two contrasted penalty functions for eliminating the diffierence (magnitude or sign) between estimators within each group.
Usage
ispls(x, y, L, mu1, mu2, eps = 1e-04, kappa = 0.05,
pen1 = "homogeneity", pen2 = "magnitude", scale.x = TRUE,
scale.y = TRUE, maxstep = 50, submaxstep = 10, trace = FALSE,
draw = FALSE)
Arguments
x |
list of data matrices, L datasets of explanatory variables. |
y |
list of data matrices, L datasets of dependent variables. |
L |
numeric, number of datasets. |
mu1 |
numeric, sparsity penalty parameter. |
mu2 |
numeric, contrasted penalty parameter. |
eps |
numeric, the threshold at which the algorithm terminates. |
kappa |
numeric, 0 < kappa < 0.5 and the parameter reduces the effect of the concave part of objective function. |
pen1 |
character, "homogeneity" or "heterogeneity" type of the sparsity structure. If not specified, the default is homogeneity. |
pen2 |
character, "magnitude" or "sign" based contrasted penalty. If not specified, the default is magnitude. |
scale.x |
character, "TRUE" or "FALSE", whether or not to scale the variables x. The default is TRUE. |
scale.y |
character, "TRUE" or "FALSE", whether or not to scale the variables y. The default is TRUE. |
maxstep |
numeric, maximum iteration steps. The default value is 50. |
submaxstep |
numeric, maximum iteration steps in the sub-iterations. The default value is 10. |
trace |
character, "TRUE" or "FALSE". If TRUE, prints out its screening results of variables. |
draw |
character, "TRUE" or "FALSE". If TRUE, plot the convergence path of loadings. |
Value
An 'ispls' object that contains the list of the following items.
x: list of data matrices, L datasets of explanatory variables with centered columns. If scale.x is TRUE, the columns of L datasets are standardized to have mean 0 and standard deviation 1.
y: list of data matrices, L datasets of dependent variables with centered columns. If scale.y is TRUE, the columns of L datasets are standardized to have mean 0 and standard deviation 1.
betahat: the estimated regression coefficients.
loading: the estimated first direction vector.
variable: the screening results of variables x.
meanx: list of numeric vectors, column mean of the original datasets x.
normx: list of numeric vectors, column standard deviation of the original datasets x.
meany: list of numeric vectors, column mean of the original datasets y.
normy: list of numeric vectors, column standard deviation of the original datasets y.
References
Liang W, Ma S, Zhang Q, et al. Integrative sparse partial least squares[J]. Statistics in Medicine, 2021, 40(9): 2239-2256.
See Also
See Also as preview.pls, ispls.cv, meta.spls, spls.
Examples
# Load a list with 3 data sets
library(iSFun)
data("simData.pls")
x <- simData.pls$x
y <- simData.pls$y
L <- length(x)
prev_pls <- preview.pls(x, y, L, scale.x = TRUE, scale.y = TRUE)
res_homo_m <- ispls(x = x, y = y, L = L, mu1 = 0.05, mu2 = 0.25,
eps = 5e-2, trace = TRUE, draw = TRUE)
res_homo_s <- ispls(x = x, y = y, L = L, mu1 = 0.05, mu2 = 0.25,
eps = 5e-2, kappa = 0.05, pen1 = "homogeneity",
pen2 = "sign", scale.x = TRUE, scale.y = TRUE,
maxstep = 50, submaxstep = 10, trace = FALSE, draw = FALSE)
res_hete_m <- ispls(x = x, y = y, L = L, mu1 = 0.05, mu2 = 0.25,
eps = 5e-2, kappa = 0.05, pen1 = "heterogeneity",
pen2 = "magnitude", scale.x = TRUE, scale.y = TRUE,
maxstep = 50, submaxstep = 10, trace = FALSE, draw = FALSE)
res_hete_s <- ispls(x = x, y = y, L = L, mu1 = 0.05, mu2 = 0.25,
eps = 5e-2, kappa = 0.05, pen1 = "heterogeneity",
pen2 = "sign", scale.x = TRUE, scale.y = TRUE,
maxstep = 50, submaxstep = 10, trace = FALSE, draw = FALSE)