Generating Synonyms from All Sources

This page describes processes to generate LexSynonyms from all sources:

• CUI - Meta-Thesaurus, for words with the same concept (CUI)
• EUI - Lexicon nominalization
• NLP_LVG - original synonyms from the Lexical Tools

I. Pre-Process

• Directory: ${LEXICON_SYNONYMS}
• program: ./Synonym/GenTaggedSPairs.java
• Inputs:
  • ./Tags/SynonymCan_Tagged.txt.${YEAR}
  • ${IN_DIR}LRNOM
  • ${IN_DIR}synonyms.data.lvg
  • ${IN_DIR}LRSPL
• Outputs:
  • ./Results/sPairsTag.data (all sPairs with tags [Y|N])
  • ./Results/sPairsTag.data.N (invalid sPairs with tag [Y|N])
  • ./Results/sPairsTag.data.Y (valid sPairs with tag [Y|N])
  • ./Results/sPairsTag.data.release (used for annual synonym.data)
• Algorithm:
  • add sPairs with tags from tagged sClasses
    • get sPairs and assign tag [Y|N]
    • sPair generation includes nom and spVar
    • if sPairs has same key ([syn1|POS1|syn2|POS2])
      • same source (CUI)
        • same tag
          => don't add (duplicates)
        • different tag
          => don't add, assign tag to [T] and ignore [N]
          => update keepTagNo, reTagno, yesNo, noNo
      • different source (CUI)
        => add sPair [syn1|POS1|syn2|POS2|src|tag]
    • assign source to CUI
  • add sPairs with tags from Lexicon nominalization
    • all nominalization are sPairs [Y]
    • assign source to EUI of noun
  • add sPairs with tags from LVG
    • all sPairs in LVG are true [Y]
    • assign source to NLP_LVG
    • removed duplicates (same key: [syn1|POS1|syn2|POS2])

II. Process

• Directory: ${LEXICON_SYNONYMS}/bin
• program: GetSynonyms ${year}

  Option	Descriptions	Inputs	Outputs
  17	Synonym.GenTaggedSPairs.java Synonym.SPairTagObj.java Synonym.SPairTagMap.java	./Tags/SynonymCan_Tagged.txt.${YEAR} ${IN_DIR}LRNOM ${IN_DIR}synonyms.data.lvg ${IN_DIR}LRSPL	./Results/sPairsTag.data => used to tag sPairs from other source (WordNet). ./Results/sPairsTag.data.N ./Results/sPairsTag.data.Y => Used for annual synonym.data release