Table 2 Model comparison between random forest and the final candidates for the logistic regression models (Land use change model)
From: Modeling land use change and forest carbon stock changes in temperate forests in the United States
Model | Accuracy | % Omission | % Commission | Class | Precision | Recall |
|---|---|---|---|---|---|---|
Random forest | 96.589 | 0.05 | 3.37 | 0 | 0 | 0 |
100 | 100 | 1 | 100 | 1 | ||
Logistic model | ||||||
 State as a fixed effect | 96.68 | 0.02 | 3.3 | 0 | 71.43 | 1.48 |
98.52 | 28.57 | 1 | 28.57 | 1 | ||
 Forest type as a fixed effect | 96.7 | 0.03 | 3.27 | 0 | 72.73 | 2.37 |
97.63 | 27.27 | 1 | 27.27 | 1 | ||
 State as a random effect | 96.68 | 0.02 | 3.3 | 0 | 71.43 | 1.48 |
98.52 | 28.57 | 1 | 28.57 | 1 | ||
 Forest type as a random effect | 96.69 | 0.03 | 3.28 | 0 | 70 | 2.08 |
97.92 | 30 | 1 | 30 | 1 | ||
 Forest type nested within state as a random effect | 96.71 | 0.03 | 3.26 | 0 | 75 | 2.67 |
97.33 | 25 | 1 | 25 | 1 | ||