addWordnet: Use WordNet to assist concept database creation

In anoopshah/Rdiagnosislist: Manipulate SNOMED CT Diagnosis Lists

Use WordNet to assist concept database creation

Description

Adds terms from a WordNet thesaurus to a concept database, matching on term. It is recommended to restrict the wordnet categories to ensure that words with multiple meanings are not linked to the wrong synonym. This function also corrects some known errors in WordNet to avoid them being passed on to the CDB; currently this applies to 'allergy = allergic reaction', 'cuneiform bone = triquetral' and 'trauma' = 'injury', but more corrections can be done if needed.

Usage

addWordnet(
  CDB_TABLE,
  wn_categories,
  WN,
  CHECK_TABLE = NULL,
  errors_to_remove = list(c("allergy", "allergic reaction"), c("allergic",
    "allergic reaction"), c("trauma", "injury"), c("traumatic", "injury"), c("skinny",
    "skin"), c("cuneiform bone", "triquetral bone"), c("upset", "disorder"),
    c("disorderliness", "disorder")),
  noisy = TRUE
)

Arguments

CDB_TABLE	data.frame or data.table with columns conceptId (integer64) and term (character, with space before and after) containing existing descriptions in the CDB
wn_categories	WordNet categories to use
WN	WordNet data.table as returned by downloadWordnet
CHECK_TABLE	other table in the same format as CDB_TABLE to check for WordNet synonyms that link to another unrelated concept, where this synonym will be excluded because of the risk of errors
errors_to_remove	list of character vectors of length two containing synonym pairs to be removed. The first entry of the pair will be removed from the WordNet file before it is used for adding to CDB
noisy	whether to output test comments

Value

CDB_TABLE with extra rows for Wordnet synonyms

References

https://wordnet.princeton.edu/

See Also

downloadWordnet

Other CDB functions: addComposeLookupToCDB(), batchDecompose(), createCDB(), downloadOrphanet(), downloadWordnet()

Examples

WORDNET <- data.table::data.table(cat = c('noun.body', 'noun.state'),
  wordnetId = bit64::as.integer64('1', '2'),
  synonyms = list(c('heart', 'pump', 'ticker'),
  c('infection', 'infectious')),
  parents = list('cardiovascular system',
  'pathologic process'), 
  adj = list('cardiac', 'infectious'))
# Add Wordnet synonyms to a concept database table
SNOMED <- sampleSNOMED()
CDB_TABLE <- description(c('Heart', 'Infection'),
  include_synonyms = TRUE)[type == 'Synonym',
  list(conceptId, term = paste0(' ', tolower(term), ' '))]
addWordnet(CDB_TABLE, 'noun.state', WORDNET)

anoopshah/Rdiagnosislist documentation built on March 29, 2025, 10:46 a.m.