Interpreting SCORE
Closed this issue · 2 comments
Is there a way to interpret the scores that I get from pymetamap
. I see that I get different scores for matches from metamap run on the command line and from the pymetamap. I assume this is because of different default settings in each case.
I am trying to see if I can set a certain threshold and filter bad matches. I dont know if this approach is correct.
Example:
text: John had a huge heart attack
metamap command line:
score=901
output:
901 -- Heart Attack (Myocardial Infarction) [Disease or Syndrome]
pymetamap:
Score=13.22
output:
ConceptMMI(index='2', mm='MMI', score='13.22', preferred_name='Myocardial Infarction', cui='C0027051', semtypes='[dsyn]', trigger='["-- Heart Attack"-tx-1-"heart attack"-noun-0]', location='TX', pos_info='17/12', tree_codes='C14.280.647.500;C14.907.585.500')
Thanks,
Deepak
The score is from metamaps MMI format. For more information about that please see the following document: https://metamap.nlm.nih.gov/Docs/MMI_Output_2016.pdf
Thanks, Anthony. The scoring process looks like pretty complicated, I have to manually look at the matches to make sure the matches are correct.
To people who visit this page later, here is the link to the algorithm: https://ii.nlm.nih.gov/MTI/Details/mmi.shtml