[Tutor] Regression analysis using statsmodels: linear combinations

[Cameron Simpson]

On 28Aug2014 17:41, Alan Gauld <alan.gauld at btinternet.com> wrote:
>On 28/08/14 12:20, Shakoor Hajat wrote:
>>I would very much appreciate any help with the following query:
>>I am running a regression analysis  in Python using statsmodels and wish
>>to obtain the combined effect of a number of the explanatory variables
>>in the model.  I could simply sum the individual coefficients to obtain
>>my effect estimate but this wouldn't work for the standard errors due to
>>covariance.  Is there a straightforward way in statsmodels to derive the
>>linear combination of coefficients?
>
>This list is for people learning the basic python language and
>standard library. This question looks a tad advanced for that
>audience. In fact I suspect that many(most?) of the folks on
>this list (including me) have only a hazy idea what you are
>talking about in your post.
>
>However there are some data analysts amongst us so they may
>be able to help.
>
>If we knew what toolkits you were using to do the analysis we
>might be able to suggest a better forum? eg numpy, RPy or pandas?
>Also what version of Python you had( 2 or 3?).

It looks like he's using this:

   http://statsmodels.sourceforge.net/

Works in python 2 and 3, but does not seem to have a dedicated mailing list.

Shakoor, if you get no responses here, ask again on python-list: I know there 
are some statistics and floating point savvy people there.

Cheers,
Cameron Simpson <cs at zip.com.au>

To be positive: To be mistaken at the top of one's voice.
Ambrose Bierce (1842-1914), U.S. author. The Devil's Dictionary (1881-1906).