# A Finnish News Corpus for Named Entity Recognition

## TL;DR

This paper introduces a Finnish news corpus with manually annotated named entities, facilitating research in Finnish NER, and evaluates baseline models on in-domain and out-of-domain datasets.

## Contribution

It provides a new Finnish news corpus with detailed annotations and baseline NER experiments, supporting future research in Finnish language processing.

## Key findings

- Baseline models achieve moderate accuracy on the corpus.
- Deep learning models outperform rule-based systems.
- Performance varies between in-domain and out-of-domain test sets.

## Abstract

We present a corpus of Finnish news articles with a manually prepared named entity annotation. The corpus consists of 953 articles (193,742 word tokens) with six named entity classes (organization, location, person, product, event, and date). The articles are extracted from the archives of Digitoday, a Finnish online technology news source. The corpus is available for research purposes. We present baseline experiments on the corpus using a rule-based and two deep learning systems on two, in-domain and out-of-domain, test sets.