CODEWITHSUNDEEP
nlpdatasetnamed-entity-recognition
DUMRE ANUP
DUMRE ANUP

Reputation: 11

Name Entity Recognition in of own language

I have a dataset of my native language, can I generate a Named Entity Recognition(NER) model for my language? How should I proceed with this, Is any tutorial based on this that develops a NER model of my own language which helps to learn from scratch?

Upvotes: 0

Views: 64

Answers (1)

Oguzhan
Oguzhan

Reputation: 214

You have several options. If you have a corpus of your language, you can train word embedding model, word2vec, and use your trained word embeddings to train sequential models, e.g., BiLSTM. Also, if you have a very large corpus, you can even pretrain a transformer-based model like BERT to fine-tune in your labeled NER dataset. If you don't have such corpus, you can use CRF model with hand-crafted features, such as length of the words and whether words are titled or not etc.

Upvotes: 1

Related Questions