irls.nbp.1: Estiamte the regression coefficients in an NBP GLM model

In gu-mi/NBGOF: Goodness-of-Fit Tests and Model Diagnostics for Negative Binomial Regression of RNA Sequencing Data

Description

Estimate the regression coefficients in an NBP GLM model for one gene

Estimate the regression coefficients in an NBP GLM model for one gene

Usage

irls.nbp.1(y, s, x, phi0, alpha1, beta0 = rep(NA, p), maxit = 50,
  tol.mu = 0.001/length(y), print.level = 1)

irls.nbp.1(y, s, x, phi0, alpha1, beta0 = rep(NA, p), maxit = 50,
  tol.mu = 0.001/length(y), print.level = 1)

Arguments

y	an n vector of counts
s	a scalar or an n vector of effective library sizes
x	a n by p design matrix
alpha1	phi= phi0 (mu/s)^alpha1
beta0	the regression coefficients: non-NA components are hypothesized values of beta, NA components are free components
tol.mu	convergence criteria
y	an n vector of counts
s	a scalar or an n vector of effective library sizes
x	a n by p design matrix
alpha1	phi= phi0 (mu/s)^alpha1
beta0	the regression coefficients: non-NA components are hypothesized values of beta, NA components are free components
tol.mu	convergence criteria

Details

This function estimate <beta> using iterative reweighted least squares (IRLS) algorithm, which is equivalent to Fisher scoring. We used the glm.fit code as a template.

Note that we will igore the dependence of the dispersion parameter (reciprical of the shape parameter) on beta. In other words, the estimate is the solution to

dl/dmu dmu/dbeta = 0

while we igored the contribution of

dl/dkappa dkappa/dbeta

to the score equation.

This function estimate <beta> using iterative reweighted least squares (IRLS) algorithm, which is equivalent to Fisher scoring. We used the glm.fit code as a template.

Note that we will igore the dependence of the dispersion parameter (reciprical of the shape parameter) on beta. In other words, the estimate is the solution to

dl/dmu dmu/dbeta = 0

while we igored the contribution of

dl/dkappa dkappa/dbeta

to the score equation.

Value

a list of the following components: beta, a p-vector of estimated regression coefficients mu, an n-vector of estimated mean values converged, logical. Was the IRLS algorithm judged to have converged? @useDynLib NBGOF Cdqrls @keywords internal

a list of the following components: beta, a p-vector of estimated regression coefficients mu, an n-vector of estimated mean values converged, logical. Was the IRLS algorithm judged to have converged? @useDynLib NBGOF Cdqrls @keywords internal

gu-mi/NBGOF documentation built on Oct. 25, 2020, 3:30 a.m.