commit | a3407ab2765080bcab4eb59c022cc3edced44e08 | [log] [tgz] |
---|---|---|
author | Marc Kupietz <kupietz@ids-mannheim.de> | Tue Aug 30 10:25:00 2022 +0200 |
committer | Marc Kupietz <kupietz@ids-mannheim.de> | Tue Aug 30 10:25:00 2022 +0200 |
tree | e677c86f1f02b6580300a5dc4b02e8382f5abfe3 | |
parent | f6e7840e09da7ed49645450a701880c8ab2c0df1 [diff] |
Improve README concerning coverage of syntagmatic analysis Change-Id: I9460cf79ebf83a394c509c909af57659b07c15d4
Julia package to load and use dereko2vec language models.
Currently, only paradigmatic analysis and count-based collocate queries are supported.
julia -e 'using Pkg; Pkg.add(url = "https://korap.ids-mannheim.de/gerrit/ids-kl/DerekoVecs.jl")'
pkg> add "https://korap.ids-mannheim.de/gerrit/ids-kl/DerekoVecs.jl"
In order to query count-based collocates, using the get_collocates()
function, collocatordb needs to be installed.
using DerekoVecs model = load("model.vecs") knn(model, "interessant", 10) cos_sim(model, "gut", "besser") model2 = load("model2.vecs") cos_sim(model, model2, "also") kld(model2, model) c = get_collocates(model)
This package is licensed under the Apache License 2.0