SteffRhes's picture
Update README.md
c1288df
metadata
language:
  - de
pretty_name: APIS ÖBL Named Entity Recognition
license: mit
task_categories:
  - token-classification

JSON file of 6,941 sentences of historical biographies, annotated with "PER" (Person), "ORG" (Organisation), "LOC" (Location).

source

The original data was extracted from the Austrian Biographical Lexicon (ÖBL) in the context of the Austrian Prosopographical Information System (APIS) project.

From there, samples were randomly pulled and annotated for Named Entity Recognition tasks, which form this dataset.

The texts concern numerous smaller biographies in the time period between 19th and early 20th century within historical Austria-Hungary, and were produced by the Austrian Acadamey of Sciences between 1957 and 2023.

The language style is rather condensed and contains a lot of domain-specific abbreviations (some of which were resolved in a related dataset: https://huggingface.co/datasets/SteffRhes/APIS_OEBL__abbreviations).

structure

json structure

The json contains a list of texts with key text_raw and the indices and types of their contained entities with key entities.

Randomized Sentences

The original data set was split into sentences and randomized samples were annotated.

no train, dev, eval split

We decided against pre-splitting the data into these sets, as their quantities might differ between requirements of various NLP training setups.

no token list

We decided against pre-tokenizing the data, as this would embed NLP logic (which tokenizer with what rule?) into the data itself.