UMLS-Core: Normalized term to CUI Mapping

• Descriptions:
  • A simple mapping from normalized term to CUI (concept)
  • File: MRXNS_ENG.RRF (fields 2 and 3)

• Examples - Test Cases:

  Input	Output	Notes
  disease's	C0012634	Genitive
  disease(s)	C0012634	parenthetical plural forms
  :disease.&*	C0012634	Punctuation
  of disease	C0012634	Stop Words
  DisEASE	C0012634	cases
  diseases	C0012634	inflectional variants
  :Disease(S).&*	C0012634	Combination
  Colours	C0009393	spelling variants
  Cortical Cystic Diseases	C0431713	Multi words

• Algorithm:
  • normalized input term by Lvg Norm
  • Mapping
    • It is one to many mapping, such as:
      • tree trumpet|C1460803
      • tree trumpet|C1492035
    • Use Java Hashtable

      	key	Values
      Type	String	Vector<String>
      Description	Normalized term	Mapped CUIs

    • Removed duplicated key|value (same normTerm|CUI with different LUI and SUI) in MRXNS_ENG.RRF.2011AB
    • MRXNS_ENG.RRF.2011AB has 9,537,805 lines. It uses lots of memory. The alternative is to use database

  • This is a one to many mapping