Combining string and phonetic similarity matching to identify misspelt names of drugs in medical records written in Portuguese

Table 3 Drugs with the highest number of misspelt variations

Drug name	Number of similar words	Inexact Phonetic Match			F1-Score when using only string match
		Precision	Recall	F1	ED≤1	ED≤2	ED≤3	ED≤4
Propanolol	52	0.960	0.979	0.967	0.310	0.819	0.945	0.955
Glibenclamida	49	1.000	1.000	1.000	0.829	0.956	0.955	0.961
Anlodipino	49	0.913	0.976	0.944	0.612	0.938	0.952	0.942
Medroxiprogesterona	47	1.000	0.914	0.955	0.763	0.881	0.955	0.927
Metoclopramida	46	1.000	0.977	0.989	0.750	0.977	0.965	0.964
Loratadina	46	0.837	0.947	0.889	0.774	0.973	0.963	0.955
Dexametasona	45	1.000	0.800	0.889	0.615	0.915	0.954	0.952
Furosemida	43	0.963	1.000	0.981	0.844	1.000	0.976	0.961
Prednisona	42	1.000	0.878	0.935	0.730	0.952	0.976	0.956
Hidroclorotiazida	41	1.000	0.975	0.987	0.776	0.962	0.952	0.940
Diclofenaco	41	0.923	0.947	0.935	0.812	0.914	0.950	0.912
Ciprofloxacino	37	1.000	0.918	0.958	0.520	0.878	0.935	0.922
Espironolactona	36	1.000	1.000	1.000	0.714	0.941	0.948	0.962
Salbutamol	36	1.000	0.972	0.986	0.819	0.956	0.976	0.943
Clonazepam	34	1.000	1.000	1.000	0.692	0.969	0.961	0.939
Beclometasona	33	1.000	0.967	0.984	0.777	0.935	0.961	0.921
Dexclorfeniramina	31	1.000	0.903	0.949	0.708	0.872	0.960	0.959
Metronidazol	30	0.965	0.965	0.965	0.816	0.964	0.942	0.926
Prednisolona	30	0.965	0.965	0.965	0.739	0.925	0.976	0.966
Isossorbida	29	0.963	1.000	0.981	0.761	0.960	0.957	0.936
Average F1-score				0.963	0.718	0.934	0.958	0.945

ISSN: 2041-1480