Multivariate statistical technique for high dimensional (biomedical) data analysis. sRDA is a a directional multivariate method to express the maximum variance in the predicted dataset by a linear combination of variables of the predictive dataset. Implemented in a partial least squares framework, for more details see sRDA on CRAN and the accompanying paper.
# from CRAN
install.packages("sRDA")Run sRDA with predefined penalization parameters.
# generate data with few highly correlated variahbles
dataXY <- generate_data(nr_LVs = 2,
n = 250,
nr_correlated_Xs = c(5,20),
nr_uncorrelated_Xs = 250,
mean_reg_weights_assoc_X =
c(0.9,0.5),
sd_reg_weights_assoc_X =
c(0.05, 0.05),
Xnoise_min = -0.3,
Xnoise_max = 0.3,
nr_correlated_Ys = c(10,15),
nr_uncorrelated_Ys = 350,
mean_reg_weights_assoc_Y =
c(0.9,0.6),
sd_reg_weights_assoc_Y =
c(0.05, 0.05),
Ynoise_min = -0.3,
Ynoise_max = 0.3)
# seperate predictor and predicted sets
X <- dataXY$X
Y <- dataXY$Y
# run sRDA
RDA.res <- sRDA(predictor = X, predicted = Y, nonzero = 5,
ridge_penalty = 1, penalization = "ust")
# check first 10 weights of X
RDA.res$ALPHA[1:10]Run sRDA with cross-validation to determine optimal penalization parameters:
RDA.res <- sRDA(predictor = X, predicted = Y, nonzero = c(5,10,15),
ridge_penalty = c(0.1,1), penalization = "enet", cross_validate = TRUE,
parallel_CV = TRUE)
# check first 10 weights of X
RDA.res$ALPHA[1:10]
# check the Ridge parameter and the number of nonzeros included in the model
RDA.res$ridge_penalty
RDA.res$nr_nonzeros
# check how much time the cross validation did take
RDA.res$CV_results$stimeObtain multiple latent variables (components):
RDA.res <- sRDA(predictor = X, predicted = Y, nonzero = c(5,10,15),
ridge_penalty = c(0.1,1), penalization = "enet", cross_validate = TRUE,
parallel_CV = TRUE, multiple_LV = TRUE, nr_LVs = 2, max_iterations = 5)
# check first 20 weights of X in first two component
RDA.res$ALPHA[[1]][1:20]
RDA.res$ALPHA[[2]][1:20]
# components are orthogonal to each other
t(RDA.res$XI[[1]]) %*% RDA.res$XI[[2]]sessionInfo()
R version 3.5.0 (2018-04-23)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Debian GNU/Linux buster/sid
Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.8.0
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.8.0
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8
[5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
attached base packages:
[1] parallel stats graphics grDevices utils datasets methods
[8] base
other attached packages:
[1] sRDA_1.0.0 mvtnorm_1.1-3 elasticnet_1.3 lars_1.3
[5] doParallel_1.0.17 iterators_1.0.14 foreach_1.5.2 Matrix_1.2-14
loaded via a namespace (and not attached):
[1] compiler_3.5.0 tools_3.5.0 codetools_0.2-15 grid_3.5.0
[5] lattice_0.20-35This package as well as the source repositories are licensed under MIT.