Could you please suggest me some good articles (original research or review) on comparative (comparison of their performance w.r.t accuracy, ranking, etc.) studies between text similarity measures like cosine similarity, BM25, Language modeling?
Text similarity analysis is hot topic in data mining, I suggest you to view articles published in high conferences and journals such KDD, PAKDD, ICDM an TKDE
You could check this paper it is available online
https://pdfs.semanticscholar.org/5b5c/a878c534aee3882a038ef9e82f46e102131b.pdf
Hi, Youcef and Sanad,
Thanks a lot for your valuable hints. Yes, Sanad, I already studied the paper you have proposed. It includes a good bunch of similarity measures and their characteristics. It's really good. However, rather than the descriptions of their intrinsic attributes, I am looking for the experimental tests/ comparisons (in terms of performance) between them. Any clue?
I would like to use the reduction methods for comparisons of text or other experimental tests. for example, PCA, tSNE, and so on.
Perhaps, you should concentrate on filtering models. Also, consider a couple of parallel corpora of a different nature.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude