Clinical Value Extraction on AHI (test)
74.7Segment-level RecallBERT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| BERTClassifier Group=Sequence models, Fine-tuning Epochs=100, Sequence Length=32, Batch Size=642021.09 | 74.7 | 88.03 | 97.05 | 95.1 | |
| Clinical BERTClassifier Group=Sequence models, Fine-tuning Epochs=100, Sequence Length=32, Batch Size=642021.09 | 73.15 | 91.4 | 97.43 | 94.76 | |
| kNNClassifier Group=Bag-of-words models, k-value=32021.09 | 67.13 | 85.34 | 86.44 | 93.57 | |
| BiLSTMClassifier Group=Sequence models, Embedding Type=Word2Vec (CBOW), Input dimension=1002021.09 | 64.54 | 98.43 | 96.37 | 94.06 | |
| Random ForestClassifier Group=Bag-of-words models2021.09 | 62.99 | 98.65 | 94.76 | 93.71 | |
| SVMClassifier Group=Bag-of-words models, Kernel=Polynomial2021.09 | 60.93 | 97.52 | 90.5 | 93.01 | |
| NaiveBayesClassifier Group=Bag-of-words models, Alpha=0.52021.09 | 55.77 | 43.67 | 91.79 | 75.87 | |
| LRClassifier Group=Bag-of-words models, Penalty=None2021.09 | 48.19 | 83.83 | 90.93 | 87.41 | |
| LASSO (L1)Classifier Group=Bag-of-words models, Penalty=L1 (λ = 0.01)2021.09 | 48.19 | 88.89 | 91.69 | 89.16 | |
| Ridge (L2)Classifier Group=Bag-of-words models, Penalty=L2 (λ = 0.01)2021.09 | 48.02 | 84.29 | 91.76 | 87.41 |