Table 5 Evaluation of text representation ensemble
From: Sentiment analysis of Indonesian datasets based on a hybrid deep-learning strategy
Model | Accuracy* | Precision* | Recall* | F-score* | AUC* |
|---|---|---|---|---|---|
BERT and RoBERTa + LSTM-attention [34] | 82.80 | 79.70 | 85.07 | 82.21 | 83.05 |
BERT + BiLSTM-TCN | 84.65 | 84.67 | 84.66 | 84.65 | 84.68 |
DistilBERT + BiLSTM-TCN | 82.28 | 84.33 | 81.09 | 82.62 | 82.44 |
Our Proposed model | 85.13 | 85.41 | 84.94 | 85.17 | 85.14 |