Table 9 Comparative results for SA of overall accuracy, precision, recall, and F1-scores
From: Social media text analytics of Malayalam–English code-mixed using deep learning
Model | Word embedding method | Precision | Recall | F1-score | Accuracy |
|---|---|---|---|---|---|
CNN | Word2Vec FastText | 0.7518 0.7458 | 0.7455 0.7329 | 0.7477 0.7374 | 0.7455 0.7329 |
LSTM | Word2Vec FastText | 0.7213 0.7392 | 0.6995 0.7381 | 0.7057 0.7372 | 0.6995 0.7381 |
GRU | Word2Vec FastText | 0.7641 0.7607 | 0.7603 0.7633 | 0.7615 0.7617 | 0.7603 0.7633 |
BiLSTM | Word2Vec FastText | 0.7247 0.7374 | 0.7225 0.7284 | 0.7226 0.7297 | 0.7225 0.7284 |
BiGRU | Word2Vec FastText | 0.7020 0.7395 | 0.6965 0.7351 | 0.6961 0.7370 | 0.6965 0.7351 |
BiLSTM + CNN | Word2Vec FastText | 0.7171 0.7380 | 0.6810 0.7396 | 0.6933 0.7356 | 0.6810 0.7396 |
BiGRU + CNN | Word2Vec FastText | 0.7112 0.7416 | 0.7106 0.7203 | 0.7080 0.7276 | 0.7106 0.7203 |
LSTM + CNN | Word2Vec FastText | 0.7135 0.7215 | 0.7151 0.7292 | 0.7124 0.7225 | 0.7151 0.7292 |
GRU + CNN | Word2Vec FastText | 0.7207 0.7514 | 0.7121 0.7255 | 0.7158 0.7336 | 0.7121 0.7255 |
Transformer based classification model | XLM-R | 0.7312 | 0.7299 | 0.7302 | 0.7299 |