[R-sig-ME] problems with allocate memory

[David Duffy]

On Fri, 23 Dec 2011, Andrew Robinson wrote:

> Hi Paul,
> 
> I apologize if this has already been discussed and rejected, but I
> also echo an earlier suggestion of Douglas's that you should think
> about working with a sample of your data instead of all of it. 
> ...
> You might even try a model-averaging approach such as
> is used in e.g. imputation; fit the  model ten times to a tenth of the
> data and average the results.

And see

http://www.bepress.com/harvardbiostat/paper33/

"We propose a computationally efficient strategy based on the 
Gauss-Seidel algorithm that iteratively fits sub-models of the GLMM to 
collapsed versions of the data. The strategy is applied to investigate the 
relationship between ischemic heart disease, socioeconomic status and 
age/gender category in New South Wales, Australia, based on outcome data 
consisting of approximately 33 million records."


-- 
| David Duffy (MBBS PhD)                                         ,-_|\
| email: davidD at qimr.edu.au  ph: INT+61+7+3362-0217 fax: -0101  /     *
| Epidemiology Unit, Queensland Institute of Medical Research   \_,-._/
| 300 Herston Rd, Brisbane, Queensland 4029, Australia  GPG 4D0B994A v