turkish-shrinked-ner

Shrinked processed version (48 entity type) of the turkish-ner. Original turkish-ner dataset: Automatically annotated Turkish corpus for named entity recognition and text categorization using large-scale gazetteers. The constructed gazetteers contains approximately 300K entities with thousands of fine-grained entity types under 25 different domains. Shrinked entity types are: academic, academic-person, aircraft, album-person, anatomy, animal, architect-person, capital, chemical, clothes, country, culture, currency, date, food, genre, government, government-person, language, location, material, measure, medical, military, military-person, nation, newspaper, organization, organization-person, person, production-art-music, production-art-music-person, quantity, religion, science, shape, ship, software, space, space-person, sport, sport-name, sport-person, structure, subject, tech, train, vehicle