GitHub - carfly/bi_ilp: Bilingual Named Entity Recognition with Integer Linear Programming

carfly / bi_ilp Public

Notifications You must be signed in to change notification settings
Fork 5
Star 9

Bilingual Named Entity Recognition with Integer Linear Programming

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
data		data
ilp		ilp
ner		ner
scripts		scripts
README		README
run.sh		run.sh

Repository files navigation

Directories
- data: training and test data
- ner: Stanford NER
  + models: model files and property files
- ilp: Bilingual ILP
- scripts: Some useful scrips

Steps (See run.sh)
- CRF-based NER
 * training: build Chinese and English NER models
 * test: predict NER results and probability of each word to a tag
- Calculate PMI
 * calculate PMI scores through the NER results
- Bilingual ILP
 * predict NER results with bilingual ILP
- Evaluation
 * evaluate the performances of CRF-based and Bi_ILP-based NER