Decompositions & Factorizations

1 Overview

Matrix decompositions and factorizations (SVD, PCA, QR, Cholesky, eigen, solve, pseudoinverse) operating on HDF5Matrix objects via standard R generics.

2 Functions

2.1 crossprod

S3 generic for . Dispatches to for objects, and to for all others.

2.2 crossprod.HDF5Matrix

crossprod.HDF5Matrix

2.3 tcrossprod

S3 generic for . Dispatches to for objects, and to for all others.

2.4 tcrossprod.HDF5Matrix

tcrossprod.HDF5Matrix

2.5 svd

S3 generic for . Dispatches to for objects, and to for all others.

2.6 svd.HDF5Matrix

svd.HDF5Matrix

2.7 prcomp.HDF5Matrix

prcomp.HDF5Matrix

2.8 eigen

Overrides to dispatch on objects. For plain R matrices the call is forwarded to .

2.9 qr

Overrides to dispatch on objects. For plain R matrices the call is forwarded to .

2.10 qr.HDF5Matrix

qr.HDF5Matrix

2.11 chol.HDF5Matrix

Computes the lower-triangular Cholesky factor L such that A = L L’. The input matrix must be square and symmetric positive-definite.

2.12 solve.HDF5Matrix

Computes the matrix inverse of a symmetric positive-definite HDF5Matrix using Cholesky decomposition + back-substitution. Equivalent to for SPD matrices.

2.13 pseudoinverse

Generic function for computing the Moore-Penrose pseudoinverse. The method computes the pseudoinverse entirely on disk using block-wise SVD; the full matrix is never loaded into RAM.

Delegates to . Result stored in the same HDF5 file under by default.