Research Article
A Selective Ensemble Classification Method Combining Mammography Images with Ultrasound Images for Breast Cancer Diagnosis
Table 6
The performance of the different selection of classifiers.
| | | Accuracy | Sensitivity | Specificity | NPV | PPV | AUC |
| C1 = Naive
Bayes-M +
KNN-U + SVM-U | 88.73% | 96.77% | 82.50% | 97.06% | 81.08% | 0.8968 |
| C2 = Naive
Bayes-U + KNN-U
+ SVM-U | 85.92% | 83.87% | 87.50% | 87.50% | 83.87% | 0.8363 |
| C3 = KNN-U +
Naive Bayes-U
+ Naive Bayes-M | 84.51% | 83.87% | 85% | 87.18% | 81.25% | 0.8242 |
| C4 = KNN-U +
SVM-U + KNN-M | 83.10% | 90.32% | 77.50% | 91.18% | 75.68% | 0.8363 |
| C5 = Naïve
Bayes-U +
SVM-U + Naive Bayes-M | 84.51% | 80.65% | 87.50% | 85.37% | 83.33% | 0.8097 |
| C6 = KNN-U +
SVM-U + SVM-M | 77.46% | 83.87% | 72.50% | 85.29% | 70.27% | 0.7782 |
| C7 = KNN-U +
Naive Bayes-M
+ KNN-M | 77.46% | 77.42% | 77.50% | 81.58% | 72.73% | 0.7500 |
| C8 = KNN-U +
Naive Bayes-U
+ KNN-M | 77.46% | 74.19% | 80% | 80% | 74.19% | 0.7468 |
| C9 = KNN-U +
Naive Bayes-M
+ SVM-M; | 74.65% | 77.42% | 72.50% | 80.56% | 68.57% | 0.7306 |
| C10 = SVM-U +
Naive Bayes-M
+ KNN-M | 74.65% | 70.97% | 77.50% | 77.50% | 70.97% | 0.7234 |
|
|
