CUR/PCA + Darrays

[emmanuellujan]

Hi @jpsamaroo, could you take a look at the following code?, what needs to be changed to use CUR and PCA with Darrays? Thanks!

using MultivariateStats
using LowRankApprox
using Distributed; addprocs(2) 
using Dagger

N, M = 1000, 200
N′ = 100 # reduced no. of features(cols)
m = rand(N, M)
Dm = distribute(m, Blocks(N÷2, M))

rows, cols = cur(Dm)

pca_model = fit(PCA, Dm', maxoutdim = N′)
m_pca = predict(m_pca_cur_model, Dm')'