To whom this may concern there is a discrepancy in the algorithm that computes
semantic similarity between two words....There are two different scores that
are returned in the Java SDK vs the Webpage application.
Webpage Score is attached in the photo.
Java SDK Scores with setMFS(false);
System.out.println("Score:
"+wordNet.compareWordRelationship("rain","precipitation",
WordNet.RelatednessMetric.WUP));
Score: 0.9523809523809523
System.out.println("Score: "+wordNet.compareWordRelationship("rain","a",
WordNet.RelatednessMetric.WUP));
Score: 0.9333333333333333
These scores are totally differ
Original issue reported on code.google.com by matth...@g.ucla.edu on 9 Jul 2014 at 5:27
Original issue reported on code.google.com by
matth...@g.ucla.edu
on 9 Jul 2014 at 5:27Attachments: