Model name | Distance recurrence | Locoregional recurrence | New primary | Residual |
---|---|---|---|---|
Particular | Value | Value | Value | Value |
No. of original samples (rows) | 311 | 311 | 311 | 152 |
Total number of independent variables (columns) | 388 | 384 | 388 | 354 |
Feature selection method used | SFFS | SFFS | SFFS | SFFS |
ML algorithm | KSVM | KSVM | KSVM | KSVM |
The minority oversampling method used | ADASYN | SMOTE | SMOTE | SMOTE |
No. of samples after oversampling | 573 | 514 | 588 | 270 |
No. of synthetic samples | 262 | 203 | 277 | 118 |
No. of features selected by SFFS | 18 | 24 | 42 | 312 |
No. of original samples for class-0 | 281 | 257 | 294 | 135 |
No. of original samples for class-1 | 30 | 54 | 17 | 17 |
Mean train accuracy | 0.99 | 0.96 | 1 | 1 |
Mean test accuracy | 0.94 | 0.73 | 0.96 | 0.91 |
Sensitivity | 0.87 | 0.83 | 0.94 | 0.89 |
Specificity | 0.96 | 0.73 | 0.98 | 1.00 |
Mean training F1 score class label 0 | 0.99 | 0.96 | 1 | 1 |
Mean training F1 score class label 1 | 0.99 | 0.97 | 1 | 1 |
Mean testing F1 score class label 0 | 0.96 | 0.78 | 0.97 | 0.93 |
Mean testing F1 score class label 1 | 0.89 | 0.64 | 0.89 | 0.87 |
Base algorithm | KSVM | KSVM | KSVM | KSVM |
Mean AUC_ROC | 0.97 | 0.73 | 0.98 | 0.99 |