soft_impute

soft_svd

sparse matrix. Both CSR <code>dgRMatrix</code> and CSC <code>dgCMatrix</code> are supported.
in case of CSR matrix we suggest to load <a href="https://github.com/dselivanov/MatrixCSR">https://github.com/dselivanov/MatrixCSR</a> package
which provides multithreaded CSR*dense matrix products (if OpenMP is supported on your platform).
On many-cores machines this reduces fitting time significantly.

maximum rank of the low-rank solution.

rank

regularization parameter for nuclear norm

lambda

maximum number of iterations of the algorithms

n_iter

convergence tolerance.
Internally we keep track relative change of frobenious norm of two consequent iterations.

convergence_tol

<a rd-options="" href="/link/svd?package=rsparse&version=0.3.1" data-mini-rdoc="rsparse::svd">svd</a> like object with <code>u, v, d</code> components to initialize algorithm.
Algorithm benefit from warm starts. <code>init</code> could be rank up <code>rank</code> of the maximum allowed rank.
If <code>init</code> has rank less than max rank it will be padded automatically.

init

<code>logical</code> whether need to make final preprocessing with SVD.
This is not necessary but cleans up rank nicely - hithly recommnded to leave it <code>TRUE</code>.

final_svd

Fit SoftImpute/SoftSVD via fast alternating least squares. Based on the
paper by Trevor Hastie, Rahul Mazumder, Jason D. Lee, Reza Zadeh
by "Matrix Completion and Low-Rank SVD via Fast Alternating Least Squares" -
<a href="https://arxiv.org/pdf/1410.2596.pdf">https://arxiv.org/pdf/1410.2596.pdf</a>

Implements several algorithms for
supervised learning on sparse data and many
matrix factorizations of sparse matrices
(with a focus on applications for recommender systems).
All algorithms work on sparse matrices. Also they extensively
use BLAS and LAPACK and parallelized with OpenMP.
Implementations are reasonably fast and nicely work with large
datasets (millions of rows and millions of columns).
List of algorithms for supervised learning:
1) Elastic net regression via Follow The Proximally-Regularized leader algorithm
2) Second order Factorization Machines via stochastic gradient descent with
adaptive learning rates. Allows to learn model parameters out-of-core.
Fast - asynchronous parallel, SIMD accelerated.
List of algorithms for matrix factorization:
1) Weighted Regularazied Matrix Factorization with Alternating Least Squares (ALS)
for implicit feedback (inculding approximate Conjugate Gradient solver).
Optional non-negativity (NNMF, non-negative matrix factorization).
2) Regularazied Matrix Factorization with ALS for explicit feedback
Optional non-negativity (NNMF, non-negative matrix factorization).
3) Fast Trunceate SVD and Soft-SVD via ALS
4) Soft-Impute via fast ALS and solution in SVD form
5) LinearFlow method which learns item-item similarity matrix from the data
Clustering:
1) kmeans from Armadillo library which provides smart (similar to kmeans++)
cluster initializations.
Misc utils/methods:
1) multithreaded `%*%` and `tcrossprod()` for `<dgRMatrix, matrix>`
2) multithreaded `%*%` and `crossprod()` for `<matrix, dgCMatrix>`

Dmitriy Selivanov

rsparse

Statistical Learning on Sparse Matrices

Drew Schmidt

Han Zhang

soft_impute function

sparse matrix. Both CSR <code>dgRMatrix</code> and CSC <code>dgCMatrix</code> are supported.
in case of CSR matrix we suggest to load <a href='https://github.com/dselivanov/MatrixCSR'>https://github.com/dselivanov/MatrixCSR</a> package
which provides multithreaded CSR*dense matrix products (if OpenMP is supported on your platform).
On many-cores machines this reduces fitting time significantly.

<a rd-options='' href='svd'>svd</a> like object with <code>u, v, d</code> components to initialize algorithm.
Algorithm benefit from warm starts. <code>init</code> could be rank up <code>rank</code> of the maximum allowed rank.
If <code>init</code> has rank less than max rank it will be padded automatically.

Fit SoftImpute/SoftSVD via fast alternating least squares. Based on the
paper by Trevor Hastie, Rahul Mazumder, Jason D. Lee, Reza Zadeh
by "Matrix Completion and Low-Rank SVD via Fast Alternating Least Squares" -
<a href='https://arxiv.org/pdf/1410.2596.pdf'>https://arxiv.org/pdf/1410.2596.pdf</a>

soft_impute: SoftImpute/SoftSVD matrix factorization

Description

Usage

Arguments

Value