Research Article
Cross-Project Defect Prediction Based on Two-Phase Feature Importance Amplification
Table 4
Average F1-measure, AUC, and MCC of TFIA using different proportions samples from target project datasets added to source project datasets.
| Proportion (%) | Average F1-measure | Average AUC | Average MCC |
| 0 | 0.352 | 0.609 | 0.221 | 10 | 0.458 | 0.668 | 0.350 | 20 | 0.575 | 0.736 | 0.475 | 30 | 0.683 | 0.804 | 0.607 | 40 | 0.790 | 0.878 | 0.739 | 50 | 0.866 | 0.928 | 0.832 | 60 | 0.921 | 0.963 | 0.902 | 70 | 0.953 | 0.981 | 0.942 | 80 | 0.969 | 0.990 | 0.962 | 90 | 0.980 | 0.995 | 0.976 |
|
|