It is, in a way, about RandomForestsClassifier's implementation: it is
designed in such a way that it adds some once-off overhead per
predict_proba, while minimising the cost per sample in a batched query. But
this is a design principle that is applied across scientific Python (and
most computational science platforms).
On 18 November 2014 22:15, Nicola Sambin <[email protected]> wrote:
> Thank you Joel and Lars,
> very kind and quick.
>
> I got the point, though it still impresses me the size of the overhead.
>
> Nicola
>
> P.S. So in the end it was not exactly about random forest implementation.
> Sorry for that.
>
>
> ------------------------------------------------------------------------------
> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> with Interactivity, Sharing, Native Excel Exports, App Integration & more
> Get technology previously reserved for billion-dollar corporations, FREE
>
> http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
