# msma: Multiblock Sparse Partial Least Squares In msma: Multiblock Sparse Multivariable Analysis

## Description

This is a function for a matrix decomposition method incorporating sparse and supervised modeling for a multiblock multivariable data analysis

## Usage

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29``` ```msma(X, ...) ## Default S3 method: msma( X, Y = NULL, Z = NULL, comp = 2, lambdaX = NULL, lambdaY = NULL, lambdaXsup = NULL, lambdaYsup = NULL, eta = 1, type = "lasso", inX = NULL, inY = NULL, inXsup = NULL, inYsup = NULL, muX = 0, muY = 0, defmethod = "canonical", scaling = TRUE, verbose = FALSE, intseed = 1, ... ) ## S3 method for class 'msma' print(x, ...) ```

## Arguments

 `X` a matrix or list of matrices indicating the explanatory variable(s). This parameter is required. `...` further arguments passed to or from other methods. `Y` a matrix or list of matrices indicating objective variable(s). This is optional. If there is no input for Y, then PCA is implemented. `Z` a vector, response variable(s) for implementing the supervised version of (multiblock) PCA or PLS. This is optional. The length of Z is the number of subjects. If there is no input for Z, then unsupervised PLS/PCA is implemented. `comp` numeric scalar for the maximum number of componets to be considered. `lambdaX` numeric vector of regularized parameters for X, with a length equal to the number of blocks. If lambdaX is omitted, no regularization is conducted. `lambdaY` numeric vector of regularized parameters for Y, with a length equal to the number of blocks. If lambdaY is omitted, no regularization is conducted. `lambdaXsup` numeric vector of regularized parameters for the super weight of X with length equal to the number of blocks. If omitted, no regularization is conducted. `lambdaYsup` numeric vector of regularized parameters for the super weight of Y with length equal to the number of blocks. If omitted, no regularization is conducted. `eta` numeric scalar indicating the parameter indexing the penalty family. This version contains only choice 1. `type` a character, indicating the penalty family. In this version, only one choice is available: "lasso." `inX` a vector or list of numeric vectors specifying the variables in X, always included in the model `inY` a vector or list of numeric vectors specifying the variables in Y, always included in the model `inXsup` a (list of) numeric vector to specify the blocks of X which are always in the model. `inYsup` a (list of) numeric vector to specify the blocks of Y which are always in the model. `muX` a numeric scalar for the weight of X for the supervised case. 0 <= muX <= 1. `muY` a numeric scalar for the weight of Y for the supervised case. 0 <= muY <= 1. `defmethod` a character representing the deflation method. This version has only the choice "canonical." `scaling` a logical, indicating whether or not data scaling is performed. The default is TRUE. `verbose` information `intseed` seed number for the random number in the parameter estimation algorithm. `x` an object of class "`msma`", usually, a result of a call to `msma`

## Details

`msma` requires at least one input X (a matrix or list). In this case, (multiblock) PCA is conducted. If Y is also specified, then a PLS is conducted using X as explanatory variables and Y as objective variables. This function scales each data matrix to a mean of 0 and variance of 1 in the default. The block structure can be represented as a list. If Z is also specified, a supervised version is implemented, and the degree is controlled by muX or muY, where 0 <= muX <= 1, 0 <= muY <= 1, and 0 <= muX + muY < 1. If a positive lambdaX or lambdaY is specified, then a sparse estimation based on the L1 penalty is implemented.

## Value

 `dmode` Which modes "PLS" or "PCA" `X` Scaled X which has a list form. `Y` Scaled Y which has a list form. `Xscale` Scaling information for X. The means and standard deviations for each block of X are returned. `Yscale` Scaling information for Y. The means and standard deviations for each block of Y are returned. `comp` the number of componets `wbX` block loading for X `sbX` block score for X `wbY` block loading for Y `sbY` block score for Y `ssX` super score for X `wsX` super loading for X `ssY` super score for Y `wsY` super loading for Y `nzwbX` number of nonzeros in block loading for X `nzwbY` number of nonzeros in block loading for Y `nzwsX` number of nonzeros in super loading for X `nzwsY` number of nonzeros in super loading for Y `selectXnames` names of selected variables for X `selectYnames` names of selected variables for Y `avX` the adjusted variance of the score for X `avY` the adjusted variance of the score for Y `cpevX` the cumulative percentage of the explained variance for X `cpevY` the cumulative percentage of the explained variance for Y `reproduct` Predictivity. Correlation between Y and the predicted Y `predictiv` Reproductivity. Correlation between the score for Y and the outcome Z

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15``` ```##### data ##### tmpdata = simdata(n = 50, rho = 0.8, Yps = c(10, 12, 15), Xps = 20, seed=1) X = tmpdata\$X; Y = tmpdata\$Y ##### One Component ##### fit1 = msma(X, Y, comp=1, lambdaX=2, lambdaY=1:3) fit1 ##### Two Component ##### fit2 = msma(X, Y, comp=2, lambdaX=2, lambdaY=1:3) fit2 ##### Sparse Principal Component Analysis ##### fit3 = msma(X, comp=5, lambdaX=2.5) summary(fit3) ```

msma documentation built on June 25, 2021, 5:09 p.m.