It's nice to see some decent speed-up factors, though the accuracy tradeoff
is still not so great. Still, I'd like to see the code and where we can go
from here. Great work so far!
On 7 August 2015 at 07:50, Maheshakya Wijewardena
wrote:
> I did a rough implementation, setting b = min_hash_match
I did a rough implementation, setting b = min_hash_match. The result I got
from running the benchmark is attached. It was able to roughly triple the
speed-up of kneighbors function for large index sizes. However, this
implementation adds some overhead to fitting time as there are 2**b *
n_estimator
Sebastian,
That does indeed help. I now understand that the calculated importance is
indeed the average Gini importance. Thank you very much!
Efrem Braun
Hi, Efrem,
I agree, this can maybe cause confusion.
However, to me,
1)
> expected fraction of samples they contribute to, (though it is not
Hi Jaret.
Please stay on the mailing list so that everybody can answer.
For 5091, as I said, best discuss it there, and if you have a question,
just post it there. People that follow github will see it.
For finding bugs: If you see an issue and it looks fixed, definitely ask
"is this fixed in
Hey Jaret.
It is usually easier to discuss these things on the github issue tracker.
Which is your pull request? Just ask there.
For the doctests you can do "make test-doc" that will run nosetests with
the appropriate options.
For the whitespace, there is an option to ignore whitespace changes.
