Skip to main content

Table 1 Evaluation results on the official test set

From: Deep learning with sentence embeddings pre-trained on biomedical corpora improves the performance of finding similar sentences in electronic medical records

 # input modelsTest set correlation
Original submissions
 Random Forest (submission #1)10.8106
 Random Forest + Dense Network (submission #2)20.8246
 Ensemble model (submission #3)80.8328
 Random Forest + Encoder Network (submission #4)20.8258
Improved models
 Random Forest10.8246
 Encoder Network10.8384
 Ensemble model30.8528