normexp {limma} | R Documentation |
Marginal log-likelihood of foreground values for the normal + exponential convolution model and its derivatives.
These functions are called by normexp.fit
and are not normally called directly by the user.
normexp.m2loglik(par,x) normexp.m2loglik.saddle(par,x) normexp.grad(par,x)
par |
numeric vector of parameters |
x |
numeric vector of (background corrected) intensities |
The parameter vector par
holds the normal mean, the normal log-standard deviation and the exponential mean.
normexp.m2loglik
computes minus twice the log-likelihood, and normexp.grad
it is derivative, based on the $normal(μ,σ^2)+exponential(α)$ convolution model for the intensities.
The elements of par
are $μ$, $log(σ)$ and $log(α)$.
normexp.m2loglik
is the saddle-point approximation to the log-logelihood, which is generally prefered because it is numerically more stable.
normexp.m2loglik
returns a numeric scalar holding minus-twice the log-likelihood.
normexp.grad
returns a numeric vector holding the derivatives with respect to the elements of par
.
Jeremy Silver and Gordon Smyth
Ritchie, M. E., Silver, J., Oshlack, A., Silver, J., Holmes, M., Diyagama, D., Holloway, A., and Smyth, G. K. (2007). A comparison of background correction methods for two-colour microarrays. Bioinformatics http://bioinformatics.oxfordjournals.org/cgi/content/abstract/btm412
An overview of background correction functions is given in 04.Background
.