Hi,
It looks to me that only the bias parameter is passed to liblinear. It is
set to self.intercept_scaling if fit_intercept is set to True and -1
otherwise. The rest (I think) are default parameters.
You can have a look at the function signature of train_wrap in
liblinear.pyx in order to clarify
Probably my question was not clear.
What I meant to ask is this: when using sklearn iblinear with default
parameters, what would be the options passed to liblinear library for training,
which is the "options" field in liblinear python wrapper?
Thank you!
From: Pagliari, Roberto [mailto:rpagli.
Dear scikit-learners,
Thank you first of all for writing such a wonderful machine learning
package for python. I've used scikit-learn quite a lot in the past and
it seemed to always would right away.
Yet, now I'm trying to get started with the BernoulliRBM in
scikit-learn, and I seem to be missin
And what about the number of trees?
On Wed, Aug 6, 2014 at 9:55 PM, Maheshakya Wijewardena <
pmaheshak...@gmail.com> wrote:
> Actually in our implementation of LSH Forest, we have an extra parameter
> to control the candidate acquisition(to avoid having the candidates with
> very small hash leng
Actually in our implementation of LSH Forest, we have an extra parameter to
control the candidate acquisition(to avoid having the candidates with very
small hash length matches - lower bound for max_depth) for `kneighbors`
queries. But that too could be controlled by some heuristic method.
But in
Lshforest as opposed to vanilla lsh, has essentially one index time
parameter: number of copies of index. It is a rather easy space,time vs
precision parameter. We could set it heuristically to increase slowly with
data dimension, so the relative overhead decreases, and then users
shouldn't really
On 6 August 2014 20:04, Lars Buitinck wrote:
> 2014-08-06 7:52 GMT+02:00 Joel Nothman :
> > Instead, could we have an interface in which the `algorithm` parameter
> could
> > take any object supporting `fit(X)`, `query(X)` and `query_radius(X)`,
> such
> > as an LSHForest instance? Indeed you cou
On 6 August 2014 20:04, Lars Buitinck wrote:
> 2014-08-06 7:52 GMT+02:00 Joel Nothman :
> > Instead, could we have an interface in which the `algorithm` parameter
> could
> > take any object supporting `fit(X)`, `query(X)` and `query_radius(X)`,
> such
> > as an LSHForest instance? Indeed you cou
Joel, I thought the interface you mentioned should accept only instances of
estimators. Sorry, My bad.
I think Lars have a good idea. Having extra parameters in the DBSCAN to use
approximate neighbors(approximate_neighbors=True) and a dict for its
parameters seems less complex and suitable at mome
2014-08-06 7:52 GMT+02:00 Joel Nothman :
> Instead, could we have an interface in which the `algorithm` parameter could
> take any object supporting `fit(X)`, `query(X)` and `query_radius(X)`, such
> as an LSHForest instance? Indeed you could also make 'lsh' an available
> algorithm using reasonabl
I don't understand the problem. The default DBSCAN will still have
algorithm='auto'.
On 6 August 2014 17:01, Maheshakya Wijewardena
wrote:
> I too considered passing the estimator instance as a parameter to DBSCAN.
> If we want to use KDTree or BallTree, NearestNeighbor instances created
> with
As far as I know, the typical idea is to keep things as readable as
possible, and only optimize the "severe/obvious" type bottlenecks (things
like memory explosions, really bad algorithmic complexity, unnecessary data
copy, etc).
I can't really comment on your "where do the bottlenecks go" questio
Hi,
As one of Maheshakya's mentors, I want to thank Arnaud, Noel and Lars
for reviewing this code! I've learned a lot from your suggestions about
scikit learn, stuff like numpy.packbits etc, and they have certainly
helped Maheshakya improve the code. I have a couple of questions, mostly
for m
I too considered passing the estimator instance as a parameter to DBSCAN.
If we want to use KDTree or BallTree, NearestNeighbor instances created
with algorithm=kdtree or ball_tree can be passed. But Robert mentioned that
it would fail the unit test cases- the base test that ensures that all
BaseEs
14 matches
Mail list logo
