Sentiment Classification on SemEval Twitter Sentiment Analysis 2015 (test)
68.8Average F1CNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CNNEmbedding=word2vec, Convolution size=[3,4,5]2017.04 | 68.8 | — | |
| CNNEmbedding=word2vec, Convolution size=[3,4,5], Fully connected layer=false2017.04 | 68.3 | — | |
| Ensemble modelcomponents=10 CNNs and 10 LSTMs2017.04 | 67.9 | — | |
| LSTMEmbedding=word2vec2017.04 | 67.7 | — | |
| CNNEmbedding=word2vec, Convolution size=[5,6,7]2017.04 | 67.6 | — | |
| LSTMEmbedding=word2vec, Fully connected layer=false2017.04 | 67.5 | — | |
| CNNEmbedding=word2vec, Convolution size=[1,2,3]2017.04 | 67.3 | — | |
| Previous best historical scores2017.04 | 67.1 | — | |
| LSTMEmbedding=fasttext2017.04 | 66.6 | — | |
| CNNEmbedding=fasttext, Convolution size=[3,4,5]2017.04 | 66.5 | — | |
| CNNEmbedding=glove, Convolution size=[3,4,5]2017.04 | 66 | — | |
| CNNEmbedding=word2vec, Convolution size=[3,4,5], Distant training=false2017.04 | 66 | — | |
| CNNEmbedding=word2vec, Convolution size=[3,4,5], Class weights=false2017.04 | 65.9 | — | |
| LSTMEmbedding=glove2017.04 | 65.8 | — | |
| LSTMEmbedding=word2vec, Class weights=false2017.04 | 65.2 | — | |
| LSTMEmbedding=word2vec, Distant training=false2017.04 | 64.7 | — | |
| Logistic regressionbaseline=1-3 grams2017.04 | 58.6 | — | |
| BBLSTM-SL2020.04 | — | 81.2 | |
| HAABSA++Method ID=Method 42020.04 | — | 81.7 | |
| LSTM+SynATT+TarRep2020.04 | — | 81.7 | |
| PRET+MULT2020.04 | — | 81.3 | |
| SentiueSemEval Winner (SW)=true2020.04 | — | 78.7 |