Sentiment Classification on Elec
6.27Error Ratetv-embeddings
Evaluation Results
| Method | Links | |
|---|---|---|
| tv-embeddingsExtra resource=Unlabeled data2015.04 | 6.27 | |
| seq2-bown-CNNvocabulary=30K most frequent words2014.12 | 7.14 | |
| CNN2015.04 | 7.14 | |
| seq2-CNNvocabulary=30K most frequent words2014.12 | 7.48 | |
| seq-CNNvocabulary=30K most frequent words2014.12 | 7.64 | |
| NB-LM bow3 (all)n-grams=3, vocabulary_size=all (up to 5M)2014.12 | 8.11 | |
| NB-LMFeatures=1-3grams2015.04 | 8.11 | |
| bow-CNNvocabulary=30K most frequent words2014.12 | 8.39 | |
| NN bow3 (all)n-grams=3, vocabulary_size=all (up to 5M)2014.12 | 8.48 | |
| dense NNFeatures=1-3grams2015.04 | 8.48 | |
| SVM bow3 (all)n-grams=3, vocabulary_size=all (up to 5M)2014.12 | 8.71 | |
| SVMFeatures=1-3grams2015.04 | 8.71 | |
| SVM bow2 (all)n-grams=2, vocabulary_size=all (up to 5M)2014.12 | 9.05 | |
| SVM bow3 (30K)n-grams=3, vocabulary_size=30K most frequent2014.12 | 9.16 | |
| SVM bow1 (all)n-grams=1, vocabulary_size=all (up to 5M)2014.12 | 11.71 |