Text Classification on MITRE ATT&CK cyber-intelligence data (test 20%)
81.95AccuracyMulticlass SGD Model
Evaluation Results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Multiclass SGD ModelVectorization=TF-IDF, Task type=Multiclass prediction, Training split=80%2026.01 | 81.95 | 77.95 | 82.72 | 89.69 | 79.03 | 73.07 | 84.24 | 80.55 | 59.57 | 71.42 | 85.61 | 76.21 | 71.3 | 63.88 | 81.17 | 58.13 | |
| GPT-4oMode=Zero-shot, Input format=Textual (sentences), Prompt=JSON format prompt2026.01 | 59 | 60 | 63.45 | 64.33 | 50.17 | 62.86 | 64.02 | 51.94 | 62.26 | 67.16 | 57.83 | 80.46 | 20.91 | 68.75 | 51.9 | 72.34 |