fac: Selecting the optimal number of latent variables.

In syspremed/PhenMAP: sBFAC - a sparse Bayesian FAC model for clustering expression data, finding associated phenotypes and molecular features.

Selecting the optimal number of latent variables.

Description

Select the optimal number of latent variables by varying the number of latent variables from fmin to fmax.

Usage

fac(Y, X, minq = 1, maxq = 2, epsilon = 0.01, plot.BIC = FALSE, printout = TRUE)

Arguments

Y	Expression data in samples by features format.
X	Design matrix from model.matrix function which defines how covariates are included into sBFAC.
minq	Minimum number of latent variables to be considered. Default is 1.
maxq	Maximum number of latent variables to be considered. Default is 2. fmax should be greater or equal to fmin.
epsilon	Assessing convergence.
plot.BIC	Plot BIC values for selecting the optimal number of latent variables.
printout	Print when the model have converged.

Details

Select the optimal number of latent variables by varying the number of latent variables from fmin to fmax.

Value

qopt	The optimal number of latent variables selected by BIC.
sig	Residual variability associated with each feature.
sig	Residual variability associated with each latent variable.
factors	Scores of observations on each latent variable.
loadings	Loadings of features on each latent variable.
coefficients	Regression coefficient for each covariate.

Author(s)

Gift Nyamundanda, Katie Eason, Pawan Poudel and Anguraj Sadanandam.

References

Nyamundanda et al (2015). A next generation tool enables identification of functional cancer subtypes with associated biological phenotypes.

Examples

## help(fac) 

syspremed/PhenMAP documentation built on April 2, 2022, 3:12 p.m.