Table 7 Comparison of evaluation metrics of various NLP algorithms with xFiTRNN.
Model | Dataset | A | Macro average P | Macro average R | Macro average F1 | Macro average AUC |
|---|---|---|---|---|---|---|
MLP/BPA19 | SST-2, SST-5 | 81.38 | 81.37 | 81.46 | 81.39 | 81.38 |
LR60 | DGAP | 85.40 | 85.12 | 85.80 | 85.49 | 86.35 |
SMO+DT16 | Airline Twitter | 89.47 | 91.60 | 89.50 | 96.30 | 91.57 |
CNN61 | StockTwits | 90.93 | 91.68 | 90.04 | 90.86 | 92.46 |
BERT15 | DJI, Financial News | 82.50 | 75.00 | 71.30 | 72.50 | 80.35 |
GPT-P414 | Forex Pair | 76.50 | 77.20 | 76.50 | 76.30 | 78.56 |
BART-large2 | Financial Phrasebank, SemEval- 2017 Task 5 | 94.70 | 95.00 | 94.50 | 94.70 | 95.37 |
FinBERT4 | Financial Phrasebank | 88.20 | 87.20 | 88.50 | 87.80 | 89.38 |
Proposed model (xFiTRNN) | Financial Phrasebank | 97.61 | 97.33 | 97.38 | 97.48 | 98.58 |
