Language Resources for Named Entity Annotation in the National Corpus of Polish
Résumé
We present the named entity annotation subtask of a project aiming at creating the National Corpus of Polish. We summarize the annotation requirements defined for this corpus, and we discuss how existing lexical resources and grammars for named entity recognition for Polish have been adapted to meet those requirements. We show detailed results of the corpus annotation using the information extraction platform SProUT. We also analyze the errors committed by our knowledge-based method and suggest its further improvements.