Table 2 Statistical results of ore-bearing sample prediction.
From: Mineral prospectivity prediction based on convolutional neural network and ensemble learning
Network model | The predicted results | ||||
|---|---|---|---|---|---|
The predicted result is the number of ore-bearing areas | The actual ore-bearing area is predicted to be ore-bearing | The unknown area is predicted to be ore-bearing | The non-ore area is predicted to be ore-bearing | ||
CNN | LeNet | 49 | 18 | 29 | 2 |
AlexNet | 58 | 26 | 30 | 2 | |
VGG 16 | 45 | 28 | 15 | 2 | |
ResNet 50 | 43 | 28 | 14 | 1 | |
Mobilenet V2 | 49 | 29 | 19 | 1 | |
VIT | 45 | 28 | 17 | 0 | |
Statistical comparison | Standard deviation | 3.83 | 2.78 | 5.89 | 0.67 |
Coefficient of variation | 0.08 | 0.11 | 0.28 | 0.50 | |
Ensemble learning | Simple average | 51 | 30 | 21 | 0 |
Majority voting | 42 | 30 | 12 | 0 | |
Plurality voting | 47 | 30 | 17 | 0 | |
Weighted average | 49 | 30 | 19 | 0 | |
Weighted voting | 49 | 30 | 19 | 0 | |
Statistical comparison | Standard deviation | 2.48 | 0 | 2.48 | 0 |
Coefficient of variation | 0.05 | 0 | 0.14 | 0 | |
