README.md

---
license: cc-by-4.0
language: da
---

This is the Kompetencer dataset created by:

```
@inproceedings{zhang-etal-2022-kompetencer,
    title = "Kompetencer: Fine-grained Skill Classification in {D}anish Job Postings via Distant Supervision and Transfer Learning",
    author = "Zhang, Mike  and
      Jensen, Kristian N{\o}rgaard  and
      Plank, Barbara",
    booktitle = "Proceedings of the Thirteenth Language Resources and Evaluation Conference",
    month = jun,
    year = "2022",
    address = "Marseille, France",
    publisher = "European Language Resources Association",
    url = "https://aclanthology.org/2022.lrec-1.46",
    pages = "436--447",
}

```

There are document delimiters indicated by `idx`.

Number of samples (sentences):
- train: 778
- dev: 346
- test: 262

Sources:
- STAR (house)

Type of tags:
- Generic BIO tags with keys `tags_skill` and `tags_knowledge`

Sample:
```
{
    "idx": 1, 
    "tokens": ["Du", "skal", "s\u00e6tte", "dagsordenen", "v\u00e6re", "v\u00e6rdiskabende", "og", "levere", "skarpt", "fagligt", "og", "strategisk", "med-", "og", "modspil", "."], 
    "tags_skill": ["O", "O", "B", "I", "B", "I", "O", "B", "I", "I", "I", "I", "I", "I", "I", "I"], 
    "tags_knowledge": ["O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O"]
}
```