getDefaultOpts: Get default options for BIBFA

Description Usage Details Value Author(s) See Also Examples

Description

A helper function that creates a list of options to be passed for CCA and GFA.

Usage

1

Details

To run the code with other option values, first run this function and then directly modify the entries before passing the list to CCA and GFA.

Value

R

The rank of hierarhical low-rank ARD prior. Possible values are all integers, including zero, and "full". When R equals "full" or R equals or is larger than the minimum value of the number of data sets and the number of latent factors, that is min(M,K), the prior corresponds to ARD prior with no low-rank structure.

lambda

The regularization parameter of the low-rank ARD model.

rotate

Whether to optimize for a linear transformation to make the variational updates less correlated.

init.tau

Initial values for the noise precision.

iter.crit

The iteration is terminated when the relative change in the lower bound for the marginal likelihood drops below this threshold.

iter.max

Maximum number of iterations.

opt.method

Which method to use for optimizing the rotation; "BFGS" or "L-BFGS".

lbfgs.factr

Optimization parameter of L-BFGS.

bfgs.crit

Optimization parameter of BFGS.

opt.iter

Number of iterations for the (L-)BFGS optimization.

addednoise

A small constant used to de-correlate latent variables of inactive components.

prior.alpha_0

Gamma prior for ARD.

prior.beta_0

Gamma prior for ARD.

prior.alpha_0t

Gamma prior for tau.

prior.beta_0t

Gamma prior for tau.

dropK

Whether to prune out empty factors from the model during inference.

low.mem

Whether to store and return the covariance matrices of W.

verbose

The amount of details printed while running CCA and GFA. 0=none, 1=medium, 2=high.

Author(s)

Seppo Virtanen, Eemeli Leppaaho and Arto Klami

See Also

CCA,GFA.

Examples

1
2
3
4
5
6
 # opts <- getDefaultOpts()  # Get the default options
 # opts$verbose <- 1         # Change some of them
 # opts$init.tau <- 10^5

 # Run the model with the new options
 # model <- CCAexperiment(Y,K,opts)

Example output



CCAGFA documentation built on May 2, 2019, 12:36 p.m.