wned icon indicating copy to clipboard operation
wned copied to clipboard

a2e index generation instruction lacks sort and aggregation

Open titsuki opened this issue 5 years ago • 0 comments

Hi, I found that the following command (See README) doesn't generate frequency column:

$ java -cp $CP ca.ualberta.entitylinking.kb.KBExtractor redirect aliasOut.txt redirectOut.txt

and the entity estimation process requires the frequency column (c.f., https://github.com/U-Alberta/wned/blob/master/src/ca/ualberta/entitylinking/cs/CandidateSelectionLucene.java#L177-L178). So, the instruction must add sort and aggregation process for alias data.

Cheers,

titsuki avatar Jun 30 '19 23:06 titsuki