search

dorianbrown / rank_bm25

A Collection of BM25 Algorithms in Python
Apache License 2.0
1.02k stars 86 forks source link

Citation missing #13

Closed ghost closed 3 years ago

ghost commented 3 years ago

Hi,

I would like to cite your software artifact in my research paper as I used it to for ranking, but I did not find a CITATION file. Do you have a preferred way to cite your work?

Best, nh

dorianbrown commented 3 years ago

Hey Nh,

Thanks for asking, to be honest I never realized people might want a way to cite it. I guess you're referencing this when you say a CITATION file? It shouldn't be too much work to get that set up, I'll look into it in the coming days.

Just out of curiosity, what kind of research paper are you writing? I originally created this repository for myself, so it's really neat to see other people getting some use out of it!

ghost commented 3 years ago

Hi dorian,

thanks for you quick reply! I was not aware of the cff format as the scientific community has given emphasis on Latex and, therefore, bibtex style citation f.e. https://github.com/explosion/spaCy/blob/master/CITATION

I developed a design science artifact which searches abstract and citation databases (scopus, springernature, etc.), merges and classifies studies in the field of applied Information Systems in the lifecycle of a tunnel. I use your library to rank (tokenized) abstracts according to queries, which consist of descriptive keywords provided by domain experts. Then, I use the received scores as feature to train models (svc vs. logistic regression vs. boosting trees) in order to make multiple binary-classifications, such as: is the study relevant or not?

I started with TF and TF-IDF, but nothing beats your bm25 implementation. Keep up the good work 👍

Best, nh

ghost commented 3 years ago

Hi dorian,

thanks a lot for the citation file!

Best, nh

dorianbrown commented 3 years ago

Happy to help! Ended up not being too hard, and now I know what that DOI number is I've been seeing anywhere :smile:.

If it's not too much work, let me know when/if your paper is published, I'd love to have a look at it!