Fast and Accurate Randomized Singular Value Decomposition Algorithms with 'PCAone'.

PCAone algorithms in R with RcppEigen!

Installation

# For the CRAN version
install.packages("pcaone")
# For the latest developing version
## devtools::install_github("Zilong-Li/PCAoneR")

Example

This is a basic example which shows you how to use pcaone:

library(pcaone)
mat <- matrix(rnorm(100*5000), 100, 5000)
res <- pcaone(mat, k = 10)
str(res)
#> List of 3
#>  $ d: num [1:10] 80.1 79.3 78.8 78.5 78.4 ...
#>  $ u: num [1:100, 1:10] -0.282 -0.106 -0.0348 -0.0219 0.0414 ...
#>  $ v: num [1:5000, 1:10] -0.01971 0.00974 -0.02306 -0.00957 0.01311 ...
#>  - attr(*, "class")= chr "pcaone"

Benchmarking

Let’s see the performance of pcaone compared to the other rsvd packages.

library(microbenchmark)
library(pcaone)
library(rsvd)
data(tiger)
timing <- microbenchmark(
    'SVD' = svd(tiger, nu=150, nv=150),
    'rSVD' = rsvd(tiger, k=150, q = 3),
    'pcaone.alg1' = pcaone(tiger, k=150, p = 3, method = "alg1"),
    'pcaone.alg2' = pcaone(tiger, k=150, p = 3, windows = 8),
    times=10)
print(timing, unit='s')
#> Unit: seconds
#>         expr       min        lq      mean    median        uq       max neval
#>          SVD 6.3386527 6.4493697 6.5878084 6.4936343 6.6752989 7.2448005    10
#>         rSVD 2.7598743 2.8006495 2.8523624 2.8390449 2.8630295 3.0286470    10
#>  pcaone.alg1 0.5111962 0.5174421 0.5360362 0.5257972 0.5529187 0.5814665    10
#>  pcaone.alg2 0.7594326 0.7668610 0.7872839 0.7833292 0.7878939 0.8441923    10

The above test is run on my MacBook Pro 2019 with processor 2.6 GHz 6-Core Intel Core i7. Note that the R is not linked to external BLAS or MKL routine. To proper benchmark the performance with single core, we can set the number of threads as one by export OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 MKL_NUM_THREADS=1.

References

• Zilong Li, Jonas Meisner, Anders Albrechtsen (2023). Fast and accurate out-of-core PCA framework for large scale biobank data

Todo

• [ ] add center and scale method.