Table 7 Statistical significance of differences in classification accuracy between triple combinations of classfication algorithms and proposed Hybrid algorithm
From: A novel hybrid model to predict concomitant diseases for Hashimoto’s thyroiditis
Classification methods | Proposed hybrid method | Z-test | \(x^2\)-test | p value |
|---|---|---|---|---|
SVM-RF-LR | Hybrid LR-KNN-MLP | 5.73 | 32.8 | \(<0\).0001 |
SVM-RF-KNN | Hybrid LR-KNN-MLP | 8.88 | 80.0 | \(<0\).0001 |
SVM-RF-MLP | Hybrid LR-KNN-MLP | 16.75 | 280.56 | \(<0\).0001 |
SVM-LR-KNN | Hybrid LR-KNN-MLP | 5.99 | 35.88 | \(<0\).0001 |
SVM-LR-MLP | Hybrid LR-KNN-MLP | 7.2 | 51.84 | \(<0\).0001 |
SVM-KNN-MLP | Hybrid LR-KNN-MLP | 12.4 | 153.76 | \(<0\).0001 |
RF-LR-KNN | Hybrid LR-KNN-MLP | 3.2 | 10.34 | \(=0\).0009 |
RF-LR-MLP | Hybrid LR-KNN-MLP | 15.04 | 225.5 | \(<0\).0001 |
RF-KNN-MLP | Hybrid LR-KNN-MLP | 8.72 | 76.03 | \(<0\).0001 |
LR-KNN-MLP | Hybrid LR-KNN-MLP | 11.02 | 121.44 | \(<0\).0001 |