Skip to content

Latest commit

 

History

History
20 lines (14 loc) · 808 Bytes

File metadata and controls

20 lines (14 loc) · 808 Bytes

Capstone_Custom_NER_Classifiers

Custom NER Classifiers

The Custom NER classifiers can be run with Java 1.8+ and Downloading the Stanford Core NLP classifiers and NER.

Setup needs to be done for environment variables in order to run the models.

The models are trained on tokens tagged using stanford NER Classifiers 3,4 and 7.

The features of individual classifiers are given below,

ICON-NER1	ICON-NER2	ICON-NER3	ICON-NER4	ICON-NER5

Classifier Name mehyog-csr-ner-model-icon.ser infinity-csr-ner-model-icon.ser dilafor-csr-ner-model-icon.ser maps-csr-ner-model-icon.ser twin-csr-ner-model-icon.ser useClassFeature TRUE TRUE TRUE TRUE TRUE useWord TRUE TRUE TRUE TRUE TRUE useNgrams TRUE TRUE TRUE TRUE TRUE maxNgramLength 6 7 8 9 10 usePrev TRUE TRUE TRUE TRUE TRUE useNExt TRUE TRUE TRUE TRUE TRUE