Skip to main content

Table 3 Comparison of the precisions, recalls and F-measures among the length-normalized edit distance, weighted edit distance using v c , and weighted edit distance using v n , when the F-measures were maximized

From: Discriminative application of string similarity methods to chemical and non-chemical names for biomedical abbreviation clustering

Method Data Precision Recall F-measure Threshold
Edit X 0.66857 0.61363 0.63992 < 0.125
Distance Y 0.46385 0.57731 0.51440 < 0.21428
Weighted X 0.69673 0.63461 0.66422 < 0.08571
ED (v c ) Y 0.51077 0.53327 0.52177 < 0.14117
Weighted X 0.61412 0.65384 0.63336 < 0.125
ED (v c ) Y 0.46225 0.60262 0.52318 < 0.19473