Research Article
MD-MinerP: Interaction Profiling Bipartite Graph Mining for Malware-Control Domain Detection
Table 2
Cross-validation results of MD-MinerP for different feature vectors of ENTN1 and ENTN2.
| Dataset | Feature vectors | Precision (%) | Recall (%) | F-measure (%) | Accuracy (%) | AUC (%) |
| ENTN1 | Process feature | 97.42 | 28.41 | 43.96 | 95.42 | 83.93 | Address feature | 100.00 | 99.72 | 99.86 | 99.98 | 99.97 | Trace feature | 97.35 | 21.27 | 34.88 | 94.98 | 79.61 | Lexical feature | 64.57 | 20.20 | 30.54 | 94.23 | 85.89 | Merged feature | 100.00 | 99.85 | 99.92 | 99.99 | 99.97 |
| ENTN2 | Process feature | 63.19 | 10.85 | 17.74 | 99.82 | 96.88 | Address feature | 100.00 | 48.05 | 63.60 | 99.90 | 95.86 | Trace feature | 84.44 | 28.90 | 42.70 | 99.86 | 75.29 | Lexical feature | 85.83 | 12.67 | 21.45 | 99.83 | 71.29 | Merged feature | 94.04 | 64.38 | 75.28 | 99.93 | 97.34 |
|
|