updateSandW

internal function for sandwich estimator

Contains functions for data analysis of Repeated measurement using GEE. Data may contain missing value in response and covariates. For parameter estimation through Fisher Scoring algorithm, Mean Score and Inverse Probability Weighted method combining with Multiple Imputation are used when there is missing value in covariates/response.
Reference for mean score method, inverse probability weighted method is
Wang et al(2007)<doi:10.1093/biostatistics/kxl024>.

Atanu Bhattacharjee

MIIPW

IPW and Mean Score Methods for Time-Course Missing Data

Bhrigu Kumar Rajbongshi

Gajendra K Vishwakarma

updateSandW function

<dl><dt>y</dt>
<dd>response value for GEE model</dd>
<dt>x</dt>
<dd>model matrix for the GEE model</dd>
<dt>vfun</dt>
<dd>variance function for the GLM</dd>
<dt>mu</dt>
<dd>mu vector for the GLM</dd>
<dt>w</dt>
<dd>weight matrix</dd>
<dt>D</dt>
<dd>derivation of the inverse link function</dd>
<dt>Ralpha</dt>
<dd>correlation matrix</dd>
<dt>beta</dt>
<dd>vector of beta value for GEE model</dd>
<dt>hessmat</dt>
<dd>hessian matrix</dd>
<dt>blockdiag</dt>
<dd>vector containing the dim of block matrix for block diagonal matrix</dd></dl>

Arguments

internal function for sandwich estimator — updateSandW

<dl>

<dt>y</dt>
<dd>response value for GEE model</dd>


<dt>x</dt>
<dd>model matrix for the GEE model</dd>


<dt>vfun</dt>
<dd>variance function for the GLM</dd>


<dt>mu</dt>
<dd>mu vector for the GLM</dd>


<dt>w</dt>
<dd>weight matrix</dd>


<dt>D</dt>
<dd>derivation of the inverse link function</dd>


<dt>Ralpha</dt>
<dd>correlation matrix</dd>


<dt>beta</dt>
<dd>vector of beta value for GEE model</dd>


<dt>hessmat</dt>
<dd>hessian matrix</dd>


<dt>blockdiag</dt>
<dd>vector containing the dim of block matrix for block diagonal matrix</dd>

</dl>

updateSandW: internal function for sandwich estimator

Description

Usage

Arguments

Details