memray / seq2seq-keyphrase

MIT License
318 stars 109 forks source link

About Micro and Macro scores #34

Closed shizhediao closed 5 years ago

shizhediao commented 5 years ago

In your paper, there is a clarification about the evaluation scores. I konw the difference between micro and macro scores. I was wondering why you choose Macro-averaged score in your paper? Are there any justifications? Thanks very much!

memray commented 5 years ago

In our case, macro-averaged is the doc-level average score. We think it makes more sense because the number of keyphrases varies doc by doc. Mix them together may lose this per-doc information. Hope this clarifies your question.

Rui