How to publish my results
If your research in the NLP Centre yields results that can be included in the infrastructure, we encourage you to add them to the Lindat repository.
Adding the result to the infrastructure
When making a deposit into the infrastructure, the recommended information to include is:
- description: include the related publication + information how to cite it (the citation + the data in Bib(La)TeX format
- licence: you can use the NLP web service or corpus licence
- affiliation: mark the relation to the NLP Centre (include the NLP Centre as the Publisher in the LINDAT repository)
- archive should contain 1) README file including the instructions, brief description, and information on how to cite the result (related paper); 2) .bib file of the related publication; 3) optionally, you can add a pdf of the related publication
When developing a new version of your result that is already included in the repository, update the information there as well (add a new version).
FAIR Data
Whenever possible, make your data follow the FAIR principles. The research infrastructure supports you in making so.
- Findable: while LINDAT supports search, it is great if you can provide rich and explicit metadata
- Accessible: use open protocols whenever possible, use authentication whenever needed
- Interoperable: prefer well-established metadata vocabularies whenever possible (e.g., Dublin Core for document description), provide links to other data if applicable (e.g. WikiData)
- Reusable: use community standards (e.g. GitLab, PapersWithCode), publish licences
Find more on FAIR principles: https://www.go-fair.org/fair-principles/
Webpage on nlp.fi.muni.cz/lindat
This section includes information on the recommended webpage content of a result published in the repository. It should include:
- description: a brief introduction and the related publication
- the tool/service: is it a demo, API, or command-line tool? How to use it?
- acknowledgments: related projects, how to cite, link to the repository
- licence
Example
A practical example is CzAccent.