Skip to main content

Table 8 NLM WSD results at different k-NN threshold levels

From: Collocation analysis for UMLS knowledge-based word sense disambiguation

   AEC MRD
   66 75 85 95 66 75 85 95
  LSC 0.7226 0.7220 0.7201 0.7082 0.6368 0.6368 0.6360 0.6360
SG Coll 0.7163 0.7038 0.7102 0.7055 0.6365 0.6365 0.6363 0.6363
  CollParser 0.7120 0.7198 0.7233 0.7055 0.6362 0.6364 0.6362 0.6356
  LSC 0.7052 0.7050 0.7110 0.7053 0.6348 0.6348 0.6344 0.6352
kAEC Coll 0.7128 0.7027 0.6992 0.7004 0.6358 0.6359 0.6347 0.6347
  CollParser 0.7118 0.7023 0.7079 0.6969 0.6406 0.6372 0.6356 0.6357
  1. Disambiguation results in terms of accuracy using the NLM WSD set. Several k-NN values are used in combination with the semantic group (SG) and the automatic extracted corpus (kAEC) methods. The disambiguation methods AEC and MRD are compared.