New Ranker with Word Mover's Distance #192
Open
+1,670
−12
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Ranker based on Word Mover's Distance that uses word embedding for calculating a distance between two documents (see Kusner, Matt, et al. "From word embeddings to document distances." 2015 - http://mkusner.github.io/publications/WMD.pdf).
config.toml example:
Available modes:
emd
- more precise according to paper above but very slowrwmd
- relaxed wmd - faster and close toemd
in precisionwcd
- word centroid discance - fast, but not preciseprefetch-prune
- combination of all modes that allow to reduce possible candidates before running emdAvailable distance functions:
cosine
- uses cosine distance between two numeric vectorsl2diff
- uses Euclidean distance (L2-norm) between two numeric vectors