Research Article
An Automatic Detection Method of Nanocomposite Film Element Based on GLCM and Adaboost M1
Table 4
The comparison of performance.
| | GABD | MLP | SVM | BayesNet | BP | RBF |
| Correctly classified instances | 97.850% | 97.833% | 97.216% | 96.366% | 97.783% | 97.817% | Incorrectly classified instances | 2.150% | 2.167% | 2.784% | 3.634% | 2.217% | 2.183% | TP rate | 0.979 | 0.978 | 0.972 | 0.964 | 0.978 | 0.978 | FP rate | 0.011 | 0.011 | 0.014 | 0.018 | 0.011 | 0.011 | Precision | 0.980 | 0.980 | 0.973 | 0.965 | 0.979 | 0.979 | Recall | 0.979 | 0.978 | 0.972 | 0.964 | 0.978 | 0.978 | -Measure | 0.978 | 0.978 | 0.972 | 0.964 | 0.978 | 0.978 | MCC | 0.968 | 0.968 | 0.959 | 0.946 | 0.967 | 0.968 | Time of building model | 0.18 s | 0.78 s | 1.34 s | 0.05 s | 5.06 s | 0.66 s |
|
|