expFarms {farms} | R Documentation |
This function converts an instance of AffyBatch
into an instance of exprSet-class
using
a factor analysis model for which a Bayesian Maximum
a Posteriori method optimizes the model parameters under the assumption
of Gaussian measurement noise.
expFarms(object, bgcorrect.method = "none", pmcorrect.method = "pmonly", normalize.method = "quantiles", weight, mu, weighted.mean, laplacian, robust, correction, centering, spuriousCorrelation, ...)
object |
An instance of |
weight |
Hyperparameter value in the range of [0,1] which determines the influence of the prior. The default value is 0.5 |
bgcorrect.method |
the name of the background adjustment method |
pmcorrect.method |
the name of the PM adjustment method |
normalize.method |
the normalization method to use |
mu |
Hyper-parameter value which allows to quantify different aspects of potential prior knowledge. Values near zero assumes that most genes do not contain a signal, and introduces a bias for loading matrix elements near zero. Default value is 0 |
weighted.mean |
Boolean flag, that indicates whether a weighted mean or a least square fit is used to summarize the loading matrix. The default value is set to FALSE. |
laplacian |
Boolean flag, indicates whether a Laplacian prior for the factor is employed or not. Default value is FALSE. |
robust |
Boolean flag, that ensures non-constant results. Default value is TRUE. |
correction |
Value that indicates whether the covariance matrix should be corrected for negative eigenvalues which might emerge from the non-negative correlation constraints or not. Default = O (means that no correction is done), 1 (minimal noise (0.0001) is added to the diagonal elements of the covariance matrix to force positive definiteness), 2 (Maximum Likelihood solution to compute the nearest positive definite matrix under the given non-negative correlation constraints of the covariance matrix) |
centering |
Indicates whether the data is "median" or "mean" centered. Default value is "median". |
spuriousCorrelation |
Numeric value in the range of [0,1] that quantifies the suppression of spurious correlation when using the Laplacian prior. Default value is 0 (no suppression). Note, that this parameter is only active when the laplacian parameter is set to TRUE. |
... |
other arguments to be passed to |
This function is a wrapper for expresso
.
data(testAffyBatch) eset <- expFarms(testAffyBatch, bgcorrect.method = "none", pmcorrect.method = "pmonly", normalize.method = "constant", weight=0.5)