Skip to main content

Table 2 Accuracy of ICD-9-CM diagnosis code prediction using large training data set (predicting top 2690 ICD-9-CM diagnosis codes having frequency at least 10 times in training data)

From: Improving medical term embeddings using UMLS Metathesaurus

Model

AUC

F1

MIC

MAC

MIC

MAC

R@8

BERT

0.9580

0.8769

0.4516

0.0932

0.3922

GloVe

0.9703

0.8888

0.4727

0.1126

0.3938

skip-gram

0.9790

0.9316

0.4995

0.1333

0.4147

fastText

0.9794

0.9340

0.4950

0.1372

0.4168

definition2vec

0.9794

0.9350

0.5065

0.1489

0.4173

  1. Bold font emphasizes the best method for each accuracy category
Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us