这是indexloc提供的服务，不要输入任何密码
Skip to content

Make LDAResults._expElogbeta() method constant #54

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 4 commits into
base: master
Choose a base branch
from
Open

Make LDAResults._expElogbeta() method constant #54

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
4aa21a6
Make LDAResults._expElogbeta() method constant
ApproximateIdentity Sep 20, 2017
63e1165
Add missing dependencies to requirements.txt
ApproximateIdentity Sep 22, 2017
8f73f70
Pin pandas to version == 0.16.2
ApproximateIdentity Oct 2, 2017
a574a0e
Remove TestLDAResults.test_expElogbeta()
ApproximateIdentity Oct 2, 2017
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 4 additions & 1 deletion requirements.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
pandas
pandas==0.16.2
scikit-learn
statsmodels
gensim
Expand All @@ -7,3 +7,6 @@ pyth
pymongo
MySQL-python
scipy
unidecode
multiprocess
nltk
10 changes: 0 additions & 10 deletions rosetta/tests/test_text.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -323,16 +323,6 @@ def test_dirichlet_expectation(self):
[-0.13470677, -13.32429878]]).T
assert_allclose(result, benchmark, atol=1e-4)

def test_expElogbeta(self):
# Make sure equal to exponential of dirichlet_expectation when we
# pass in all ones
lda = self.choose_lda('lda')
lda._lambda_word_sums = pd.Series(
np.ones(lda.num_topics), index=lda.topics)
result = lda._expElogbeta
benchmark = np.exp(lda._dirichlet_expectation(lda.pr_token_topic))
assert_frame_equal(result, benchmark)

def test_predict_1(self):
# Use fact that w0 <--> topic_0, w1 <--> topic_1
lda = self.choose_lda('lda_2')
Expand Down
7 changes: 4 additions & 3 deletions rosetta/text/vw_helpers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -332,6 +332,9 @@ def _set_probabilities(self, topics, predictions):
self.pr_doc = doc_sums / doc_sums.sum()
self.pr_doc_topic = predictions / predictions.sum().sum()

lam = self._lambda_word_sums * self.pr_token_topic
self._constExpElogbeta = np.exp(self._dirichlet_expectation(lam + EPS))

def prob_token_topic(self, token=None, topic=None, c_token=None,
c_topic=None):
"""
Expand Down Expand Up @@ -565,9 +568,7 @@ def _expElogbeta(self):
topic-word weights.
"""
# Get lambda, the dirichlet parameter originally returned by VW.
lam = self._lambda_word_sums * self.pr_token_topic

return np.exp(self._dirichlet_expectation(lam + EPS))
return self._constExpElogbeta

def _dirichlet_expectation(self, alpha):
"""
Expand Down