nbreg.fit: Negative Binomial regression.

In mstrazar/csDEX: Condition-specific differential exon expression.

Description

Fit a Negative Binomial regression model. Called internally by the geneModel function.

Usage

    nbreg.fit(X, y, phi, beta.init = NULL, offset = 0, max.iter = 100, tol = 1e-06,
    lambda.0 = 1e-04, lambda.max = 100, verbose = FALSE)

Arguments

X	The design matrix with n rows and p columns.
y	A vector of n target values.
phi	Precision hyperparameter. Either a scalar or a vector of n values.
beta.init	Optional. Initial estimate of parameter values.
offset	Optional. Known offset of target values from a negative binomially distributed data.
max.iter	Number of optimization iteration.
tol	Relative change in log-likelihood to declare convergence.
lambda.0	Initial Levenberg-Marquardt damping factor.
lambda.max	Maximal Levenberg-Marquardt damping factor.
verbose	Print diagnostic messages.

Details

Custom implementation of the Negative Binomial regression model fitting with Levenberg-Marquardt update, enabling custom initialization values.

Value

coefficients	Parameter values for the mean of each fitted value.
fitted.values	Means of n fitted values.
converged	Binary convergence factor. If FALSE, use verbose=TRUE to determine possible reasons.
phi	Precision.
loglik	Log-likelihood of model parameters given data.
iter	Number of executed gradient steps.
vcov	Parameter variance-covariance matrix.

Author(s)

Martin Stra<c5><be>ar

mstrazar/csDEX documentation built on May 23, 2019, 8:16 a.m.