Sentiment Classification on MR (10-fold cross-validation)
83.1AccuracyAdaSent
Evaluation Results
| Method | Links | |
|---|---|---|
| AdaSenttype=supervised compositional model2015.06 | 83.1 | |
| BRNNtype=supervised compositional model2015.06 | 82.3 | |
| CNNtype=supervised compositional model2015.06 | 81.5 | |
| combine-skip + NBclassifier=linear model, augmentation=bigram Naive Bayes features2015.06 | 80.4 | |
| NB-SVMtype=bag-of-words model2015.06 | 79.4 | |
| MNBtype=bag-of-words model2015.06 | 79 | |
| cBoWtype=bag-of-words model2015.06 | 77.2 | |
| RNNtype=supervised compositional model2015.06 | 77.2 | |
| combine-skipclassifier=logistic regression2015.06 | 76.5 | |
| GrConvtype=supervised compositional model2015.06 | 76.3 | |
| uni-skipclassifier=logistic regression2015.06 | 75.5 | |
| Paragraph-vectortype=unsupervised learning2015.06 | 74.8 | |
| bi-skipclassifier=logistic regression2015.06 | 73.9 |