Removing and Adding new entity labels to en_core_web_lg

[aniyyanz08]

I found that en_core_web_lg has named recognition entity labels as CARDINAL, DATE, EVENT, FAC, GPE, LANGUAGE, LAW, LOC, MONEY, NORP, ORDINAL, ORG, PERCENT, PERSON, PRODUCT, QUANTITY, TIME, WORK_OF_ART.

Out of these labels i need DATE, EVENT, ORG, PERSON, TIME, WORK_OF_ART. Other entity labels from the above are not needed for my application. Also i need few other entity labels to be included to this. Is it possible to remove few of the existing labels and add few other new labels to this list.

[ljvmiranda921]

Hi, @aniyyanz08

Out of these labels i need DATE, EVENT, ORG, PERSON, TIME, WORK_OF_ART. Other entity labels from the above are not needed for my application.

You can safely ignore or filter the labels you don't need:

doc = nlp(SOME_TEXT)
ents = [ent for ent in list(doc.ents) if ent.label_ not in LIST_OF_FILTERS]

Also i need few other entity labels to be included to this.

You can check the multiple NER demo on how to do that. From here you can combine multiple components to include your new labels