In CHuanSite/SJD: Structured Joint Decomposition (SJD)

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Below is a demo of installing and running SJD package on simulated datasets. For analytical vignette using real data, please refer to Real Data Analysis here.

Install and load `SJD` package

To install this package in R, run the following commands:

library(devtools)
install_github("CHuanSite/SJD")

library(SJD)

Simple example

# Simulation the dataset
dataset = list(matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50))

## Specify the structure among the datasets
group = list(c(1,2,3,4), c(1,2), c(3,4), c(1,3), c(2,4), c(1), c(2), c(3), c(4))
comp_num = c(2,2,2,2,2,2,2,2,2)

## Separate PCA, ICA, NMF
sepPCA_res = sepPCA(dataset, comp_num)
sepICA_res = sepICA(dataset, comp_num)
sepNMF_res = sepNMF(dataset, comp_num)

## Concatenated PCA, ICA, NMF
concatPCA_res = concatPCA(dataset, group, comp_num)
concatICA_res = concatICA(dataset, group, comp_num)
concatNMF_res = concatNMF(dataset, group, comp_num)

## Joint PCA, ICA, NMF
jointPCA_res = jointPCA(dataset, group, comp_num)
jointICA_res = jointICA(dataset, group, comp_num)
jointNMF_res = jointNMF(dataset, group, comp_num)

## twoStageLCA
twoStageLCA_res = twoStageLCA(dataset, group, comp_num)

To access the component

str(concatPCA_res$linked_component_list)

To access the score

str(concatPCA_res$score_list)

Weighting data sets by different weights

SJD package also allows users to specify weights for datasets during the analysis

# Simulation the dataset
dataset = list(matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50))

## Specify the structure among the datasets
group = list(c(1,2,3,4), c(1,2), c(3,4), c(1,3), c(2,4), c(1), c(2), c(3), c(4))
comp_num = c(2,2,2,2,2,2,2,2,2)
weighting = c(2, 1, 4, 3)

## Separate PCA, ICA, NMF
sepPCA_res = sepPCA(dataset, comp_num, weighting)
sepICA_res = sepICA(dataset, comp_num, weighting)
sepNMF_res = sepNMF(dataset, comp_num, weighting)

## Concatenated PCA, ICA, NMF
concatPCA_res = concatPCA(dataset, group, comp_num, weighting)
concatICA_res = concatICA(dataset, group, comp_num, weighting)
concatNMF_res = concatNMF(dataset, group, comp_num, weighting)

## Joint PCA, ICA, NMF
jointPCA_res = jointPCA(dataset, group, comp_num, weighting)
jointICA_res = jointICA(dataset, group, comp_num, weighting)
jointNMF_res = jointNMF(dataset, group, comp_num, weighting)

## twoStageLCA
twoStageLCA_res = twoStageLCA(dataset, group, comp_num, weighting)

Projecting new data sets to extracted components

# Simulation the dataset
dataset = list(matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50),
               matrix(runif(5000, 1, 2), nrow = 100, ncol = 50))

## Specify the structure among the datasets
group = list(c(1,2,3,4), c(1,2), c(3,4), c(1,3), c(2,4), c(1), c(2), c(3), c(4))
comp_num = c(2,2,2,2,2,2,2,2,2)

## Projected new data sets
proj_dataset = list(matrix(runif(5000, 1, 2), nrow = 100, ncol = 50))
proj_group = list(c(TRUE, TRUE, FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, FALSE))

## concatenate PCA with projection functionality on
res_concatPCA = concatPCA(dataset, group, comp_num, weighting = NULL, proj_dataset = proj_dataset, proj_group = proj_group)