Named Entity Recognition

Model Architecture:

• Pretrained Word Embedding + BiLSTM + CRF

Library:

• pytorch
• tqdm
• numpy
• gensim

Data :

• VLSP 2016

Kết quả thử nghiệm:

Entity	precision	recall	F1-score
LOC	0.88	0.91	0.89
MISC	0.94	0.94	0.94
ORG	0.75	0.58	0.65
PER	0.87	0.81	0.84
---	---	---	---
micro avg	0.87	0.82	0.84
macro avg	0.86	0.81	0.83
weighted avg	0.86	0.82	0.84