A vector of regression coefficients (e.g., a row from the coefficient matrix produced by <code><a rd-options="" href="/link/geeboost?package=threeboost&version=1.1" data-mini-rdoc="threeboost::geeboost">geeboost</a></code>)

Version of QIC to implement, for either <code>"gaussian"</code>, <code>"binomial"</code> or <code>"poisson"</code> outcomes.
Should match the <code>family</code> argument used in the original boosting algorithm.

family


Calculates a simple version of Pan's QIC for a GEE model defined by
 a vector of regression coefficients.


This package implements a thresholded version of the EEBoost
algorithm described in [Wolfson (2011, JASA)]. EEBoost is a general-purpose
method for variable selection which can be applied whenever inference would
be based on an estimating equation. The package currently implements
variable selection based on the Generalized Estimating Equations, but can
also accommodate user-provided estimating functions. Thresholded EEBoost is
a generalization which allows multiple variables to enter the model at each
boosting step.

QIC: Pan's QIC

Description

Usage

Arguments