Table 4 Performance comparison of various base, ensemble and hybrid landslide prediction models on training dataset using statistical parameters based on confusion matrix.
From: Hybrid machine learning approach for landslide prediction, Uttarakhand, India
Performance Measures | Results on Training Dataset | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
BN | HBNRS | BPNN | HBPNNRS | RF | HRFRS | Bagging | HBRS | XGBoost | HXGBRS | |
Sensitivity | 0.779 | 0.792 | 0.790 | 0.949 | 0.827 | 0.992 | 0.821 | 0.992 | 0.855 | 0.992 |
Specificity | 0.673 | 0.747 | 0.700 | 0.890 | 0.803 | 0.964 | 0.757 | 0.971 | 0.927 | 0.927 |
Precision | 0.885 | 0.903 | 0.892 | 0.946 | 0.928 | 0.982 | 0.906 | 0.985 | 0.974 | 0.989 |
F1-Score | 0.861 | 0.850 | 0.837 | 0.947 | 0.874 | 0.986 | 0.861 | 0.988 | 0.881 | 0.990 |
AUC | 0.749 | 0.897 | 0.751 | 0.977 | 0.841 | 0.989 | 0.882 | 0.995 | 0.921 | 0.996 |
Accuracy (%) | 75.42 | 80.00 | 76.86 | 93.00 | 82.00 | 98.31 | 80.48 | 98.55 | 87.71 | 98.79 |
