EMupdate: Utility function performing EM algorithm updates

In gitlongor/hisemi: Hierarchical Semiparametric Regression of Test Statistics

Description

Utility function performing EM algorithm updates for penLik.EMNewton

Usage

EMupdate(starts, nLogLik.pen, optim.method, H, tstat, df, dt0, 
        spar.Pen.mat, em.iter.max = 10, em.beta.iter.max = 1, 
        scale.conv = 0.001, lfdr.conv = 0.001, 
        NPLL.conv = 0.001, debugging = FALSE)

Arguments

starts	A numeric vector of starting values, in 'r' parameterization of the scale.fact
nLogLik.pen	A function computing negative penalized log likelihod
optim.method	One of BFGS, CG, L-BFGS-B, Nelder-Mead, SANN, nlminb, NR, the method used for optimization.
H	Design matrix
tstat	A numeric vector of t-statistics
df	A numeric scalar or vector of degrees of freedom
dt0	A numeric vector of the central t-density evaluated at the t-statistics
spar.Pen.mat	smoothing parameter times the penalty matrix
em.iter.max	Maximum number of EM iterations
em.beta.iter.max	Maximum number of iterations in maximization step with respect to regression coefficients
scale.conv	A small numeric scalar specifying the convergence criterion for the scale parameter
lfdr.conv	A small numeric scalar specifying the convergence criterion for the local false discovery rates
NPLL.conv	A small numeric scalar specifying the convergence criterion for the negative penalized log likelihood
debugging	A logical scalar indicating whether debugging mode of the code should be run

Value

A numeric vector of updated parameter estimates. The scale factor is in the log(scale.fact-1) parameterization.

Author(s)

Long Qu [e-mail]

References

Long Qu, Dan Nettleton, Jack Dekkers (2012) A hierarchical semiparametric model for incorporating inter-gene relationship information for analysis of genomic data. Biometrics, 68(4):1168-1177

See Also

penLik.EMNewton, NRupdate

gitlongor/hisemi documentation built on May 17, 2019, 5:28 a.m.