deep learning/results with TRAIN.md

LSTM

• Precision: 0.8627
• Recall: 0.3557
• F1: 0.2897
• Accuracy: 0.5924
• Confusion matrix: [[3, 213, 0], [0, 431, 0], [0, 89, 5]]

Full classification report: precision recall f1-score support

positive     1.0000    0.0139    0.0274       216
 neutral     0.5880    1.0000    0.7405       431
negative     1.0000    0.0532    0.1010        94

accuracy                         0.5924       741

macro avg 0.8627 0.3557 0.2897 741 weighted avg 0.7604 0.5924 0.4515 741

GRU

• Precision: 0.8623
• Recall: 0.8200
• F1: 0.8387
• Accuracy: 0.8758
• Confusion matrix: [[179, 32, 5], [17, 405, 9], [7, 22, 65]]

Full classification report: precision recall f1-score support

positive     0.8818    0.8287    0.8544       216
 neutral     0.8824    0.9397    0.9101       431
negative     0.8228    0.6915    0.7514        94

accuracy                         0.8758       741

macro avg 0.8623 0.8200 0.8387 741 weighted avg 0.8746 0.8758 0.8737 741

CNN

• Precision: 0.9147
• Recall: 0.8296
• F1: 0.8632
• Accuracy: 0.8961
• Confusion matrix: [[180, 35, 1], [9, 420, 2], [10, 20, 64]]

Full classification report: precision recall f1-score support

positive     0.9045    0.8333    0.8675       216
 neutral     0.8842    0.9745    0.9272       431
negative     0.9552    0.6809    0.7950        94

accuracy                         0.8961       741

macro avg 0.9147 0.8296 0.8632 741 weighted avg 0.8991 0.8961 0.8930 741