dselivanov / text2vec

Fast vectorization, topic modeling, distances and GloVe word embeddings in R.
http://text2vec.org
Other
850 stars 135 forks source link

jspca_robust #274

Closed manuelbickel closed 4 years ago

manuelbickel commented 6 years ago

With reference to #233 I have set up a fixed version of the jsPCA function that could be fed into the createJSON call, so that no large changes would be necessary to existing text2vec code. NOTE: I have not invented the fix, details on the source are inlcuded in the code.

Of course, this PR does not solve #233, but only a little part of it.

In case you think this PR makes sense, please let me know if you desire any changes or if I have overseen something codewise. If you feel that the PR is does not make sense, we can simply delete/close it.

codecov[bot] commented 6 years ago

Codecov Report

Merging #274 into master will decrease coverage by 0.26%. The diff coverage is 0%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #274      +/-   ##
==========================================
- Coverage   78.94%   78.68%   -0.27%     
==========================================
  Files          39       39              
  Lines        2394     2402       +8     
==========================================
  Hits         1890     1890              
- Misses        504      512       +8
Impacted Files Coverage Δ
R/model_LDA.R 73.68% <0%> (-0.56%) :arrow_down:
R/utils.R 46.42% <0%> (-15.48%) :arrow_down:

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 718dac7...0c82a1a. Read the comment docs.

dselivanov commented 6 years ago

As usual I travel when you send PR :-))

manuelbickel commented 6 years ago

Just came across this during my work and wanted to bring it on the table, looking forward to discuss this when you are back. Have a good journey!

Von: Dmitriy Selivanov [mailto:notifications@github.com] Gesendet: Freitag, 27. Juli 2018 20:50 An: dselivanov/text2vec text2vec@noreply.github.com Cc: manuelbickel manuel.bickel@posteo.de; Author author@noreply.github.com Betreff: Re: [dselivanov/text2vec] jspca_robust (#274)

As usual I travel when you send PR :-))

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/dselivanov/text2vec/pull/274#issuecomment-408507609 , or mute the thread https://github.com/notifications/unsubscribe-auth/ASaYqnbO8wZ577pfOVG8Oii1FGWW5gL_ks5uK2DNgaJpZM4Vgs6N . https://github.com/notifications/beacon/ASaYqlu-mxkK41-ioWcXH-LtqCkp4NMpks5uK2DNgaJpZM4Vgs6N.gif