Table 2 Accuracy (%) of RQE methods using the respective training and test sets of four datasets: SNLI, MultiNLI, Quora, and Clinical-QE
Methods | Textual Datasets | Question Datasets | ||
|---|---|---|---|---|
SNLI | MultiNLI | Quora | Clinical-QE | |
Neural Network (NN) | 79.50 | 73.71 | 81.34 | 71.45 |
NN + GloVe embeddings | 82.80 | 78.52 | 83.62 | 93.12 |
Logistic Regression + Features | 75.91 | 67.88 | 67.79 | 98.60 |