gram: Function gram
In kperrakis/sbr: Sparse and Scalable Bayesian Regression for Linear Models

Description Usage Arguments Value Author(s) References See Also Examples

Function for calculating the (inner-product) Gram matrix that allows for block-matrix multiplication.

1	gram(X, trX, block = FALSE, block.size = 1000, show.progress = FALSE)

`X`	a standardized feature matrix.
`trX`	(optional) the transpose matrix of X.
`block`	logical, block matrix computation is performed when TRUE, default option is FALSE.
`block.size`	used when block = TRUE. Default option is 1000, i.e. blocks of dimensionality 1000 x 1000.
`show.progress`	logical, when TRUE (and block = TRUE) the progress of the calculations is reported on the console.

Returns the inner-product Gram matrix.

Konstanstinos Perrakis konstantinos.perrakis@dzne.de

Sach Mukherjee sach.mukherjee@dzne.de

Perrakis, K., Mukherjee, S. and the Alzheimers Disease Neuroimaging Initiative. (2018) Scalable Bayesian regression in high dimensions with multiple data sources. https://arxiv.org/pdf/1710.00596.pdf

gram.parallel

X <- matrix(rnorm(100 * 300), 100, 300)
G0 <- gram(X)                             # usual matrix multiplication
G1 <- gram(X, block=TRUE,block.size=100)  # block matrix multiplication
all.equal(G0, G1)