metadata
language:
- en
metrics:
- accuracy
pipeline_tag: text-classification
datasets:
- AliArshad/Bugzilla_Eclipse_Bug_Reports_Dataset
Model Training Summary
Data Splitting
- Split Ratio: 80:20 (Training:Test)
Label Mapping
label_mapping = {'non-severe': 0, 'severe': 1}
Training Loss Over Iterations
| Step | Training Loss |
|---|---|
| 500 | 0.518800 |
| 1000 | 0.321000 |
| 1500 | 0.210000 |
| 2000 | 0.120900 |
Evaluation Metrics
- Accuracy: 0.8452921586436573
- Precision: 0.8686567164179104
- Recall: 0.909375
- F1-score: 0.8885496183206106
- MCC (Matthews Correlation Coefficient): 0.637922907223529
Confusion Matrix
[[ 755 308]
[ 203 2037]]