Exclusive Filter: A Term Ends with Parenthetic Abbreviation

• Description:
  If a term ends with parenthetic abbrevication is not a valid multiword. These terms are filtered out from the MEDLINE n-gram set. For examples, the following terms are invalid multiwords:
  • activating transcription factor 3 (ATF3)
  • acute coronary syndrome (ACS)
  • angiotensin-converting enzyme (ACE)
  • Alzheimer's disease (AD)

• Filter Algorithm:
  • Logics:

    Description	FilterType	Notes
    Get the end-term	FT_TBD	Aaa bbb bbc (ABB): => (ABB):
    Remove ,.: from end-term if it exist	FT_TBD	(ABB): => (ABB)
    Check if match pattern of (ABB) end-term starts with ( end-term ends with ) abbreviation are all upper case	FT_TBD
    Check if an abbreviation The 1st character of input is the same as the 2nd character of end-term (ABB)	FT_END_TERM_NOT_ABB	Exceptions: valid terms end with pattern of (ABB), but ABB is not an abbreviation.
    Check if not an abbreviation	FT_END_TERM_INV_ABB	Filtered invalid terms ends with (ABB)

  • source code: FilterEndTermAbb.java
  • FilterType: FilterType.FT_END_TERM_INV_ABB

• Accuracy Test on Lexicon:
  • InFile:
    • ${OUT_DATA}/03.LeadEndTerm/lexWords.data
  • Result:

    Lexicon	Filter	Sample No	Pass No	Trap No	Exp No	Pass-Rate
    2016	FT_END_TERM_INV_ABB	915583	915583	0	18	100.0000%
    2014	FT_END_TERM_INV_ABB	875090	875090	0	10	100.0000%
    2015	FT_END_TERM_INV_ABB	896213	896213	0	15	100.0000%