Natural Language Processing Algorithms for Normalizing Expressions of Synonymous Symptoms in Traditional Chinese Medicine

Table 8

The top 5 and 10 recall of models.

Model	Top 5 recall	Top 10 recall
Jaccard similarity	0.57312	60857
Word2Vec with cosine	0.9145 ± 0.0025	0.9524 ± 0.0035
DNorm	0.9702 ± 0.0006	0.9858 ± 0.0008
BERT-based ranking	0.9852 ± 0.0049	0.9910 ± 0.0050
BERT-Classification	0.9864 ± 0.0044	0.9921 ± 0.0048

Note. The test results are expressed as mean ± SD. Each model was repeated 10 times, except for Jaccard similarity. , compared with BERT-Classification.