Skip to content
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.

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

Add more features to the model in vw_ensemble #288

Merged
merged 4 commits into from
Jul 2, 2019
Merged

Conversation

osma
Copy link
Member

@osma osma commented Jul 2, 2019

Related to #235

This PR adds more features to the ML models built in the vw_ensemble backend:

  • raw, descriptor invariant scores returned by the source project
  • "zero features" whenever a source project returns a zero or near-zero score for a concept
  • combinations (interactions) between raw scores

In my experiments, these additional features seem to give a small improvement in results.

@osma osma added this to the v0.41 milestone Jul 2, 2019
@codecov
Copy link

codecov bot commented Jul 2, 2019

Codecov Report

Merging #288 into master will increase coverage by <.01%.
The diff coverage is 100%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #288      +/-   ##
==========================================
+ Coverage   99.36%   99.36%   +<.01%     
==========================================
  Files          55       55              
  Lines        2842     2845       +3     
==========================================
+ Hits         2824     2827       +3     
  Misses         18       18
Impacted Files Coverage Δ
tests/test_backend_vw_ensemble.py 100% <100%> (ø) ⬆️
annif/backend/vw_ensemble.py 100% <100%> (ø) ⬆️

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 8521652...d531b74. Read the comment docs.

@osma osma merged commit 2ea1130 into master Jul 2, 2019
osma added a commit that referenced this pull request Jul 5, 2019
…they

introduce extra complexity without significantly improving results.
Reverting the merge for now. The vw_ensemble backend needs more work, but
it's better to have a simple version in master and build upon that.

Revert "Merge pull request #288 from NatLibFi/issue235-more-features"

This reverts commit 2ea1130, reversing
changes made to 8521652.
osma added a commit that referenced this pull request Jul 5, 2019
…eatures

Revert PR #288 additional features for vw_ensemble
@osma osma deleted the issue235-more-features branch September 30, 2019 12:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant