Commit Graph

17 Commits

Author SHA1 Message Date
Joshua Loyal 60b41a8044 correct optional arguments for FeatureHasher 2015-10-08 11:58:09 -04:00
Lars Buitinck 1f3ebb02fc DOC FeatureHasher takes (finite) numbers as values
Fixes #4882.
2015-07-12 13:20:06 +02:00
pianomania 243983daaf try to mollify travis 2015-07-09 20:31:02 +08:00
pianomania bf8695451b add an example for FeatureHasher 2015-07-08 21:37:07 +08:00
pianomania 4491d10ced add a example for FeatureHasher 2015-07-08 13:57:21 +08:00
Andreas Mueller 0650d5502e DOC adding backlinks to docstrings 2015-06-03 00:24:04 -04:00
Raghav R V b3fbccca38 FIX various mismatch between docstring and signature params
DOC max_iterations -> max_iter. Make it consistent with kmeans

MAINT Replace the deprecated dx parameter with d in the docstrings

MAINT Deprecation warning for max_iterations parameter.
2015-01-16 11:54:57 -05:00
Lars Buitinck 6713132da5 DOC let OneHotEncoder, DictVectorizer and FeatureHasher refer to each other
Judging by ML and SO questions, users have a hard time finding the right
vectorizers and don't know how to trade off the benefits of these three
classes.
2013-07-18 22:35:05 +02:00
Robert Layton dacfd8bd5d DOC: Replaced all BSD style licenses with "BSD 3 clause"
Replaced all BSD style licenses with "BSD 3 clause"
Not checked yet!

Removed duplicate "3 clause, 3 clause"

Removed trailing period if exists

Fixed some missed licences, still about 50 to do, but those can be automated

Think I got the last of them.

Apparently me and sed have different ideas of regex.

Found a few more
2013-04-30 08:34:46 +02:00
Lars Buitinck 164d36f8de DOC distributions produced by hashing trick depend on input
Not guaranteed to be multinomials or Gaussians, AFAICT.
2013-04-16 13:31:52 +02:00
Andreas Mueller 9603b070d7 COSMIT pep8 2012-12-02 11:17:17 +00:00
Lars Buitinck f93ebb47e7 BUG enforce and document max. n_features for FeatureHasher
Using a larger number might work sometimes, but produce bogus
results, or it may cause cryptic error messages from scipy.sparse.

Also justify the use of a single hash function for those who have
read Weinberger's paper.
2012-11-19 12:04:57 +01:00
Lars Buitinck 3de0f43e40 ENH add dict support to FeatureHasher and make it the default input_type 2012-11-17 17:29:20 +01:00
Lars Buitinck 205a9477ca ENH + DOC FeatureHasher
Suggestions from @amueller and @GaelVaroquaux:
* default value for n_features
* input validation in fit, with test
* lots of small docfixes
* capitalization and full stop in error messages
2012-11-17 17:29:20 +01:00
Lars Buitinck cbb9b0181a ENH allow (feature, value) pairs in FeatureHasher 2012-11-17 17:29:20 +01:00
Lars Buitinck f5a4ad2bfc ENH speed up hashing and reduce memory usage by 1/3
Build CSR instead of COO matrix in FeatureHasher.
Partially followed Jochen Wersdörfer's (@ephes's) suggestions.
2012-11-17 17:29:20 +01:00
Lars Buitinck 2b8ab3561a ENH feature hashing transformer 2012-11-17 17:29:20 +01:00