Table 3 The rank of variable importance in the random forest algorithm
From: Machine learning approach to drivers of bank lending: evidence from an emerging economy
Rank | Variable | %IncMSE | IncNodePurity | Average |
|---|---|---|---|---|
1 | CAPR | 11.621 | 14.633 | 13.127 |
2 | DEPOVER | 10.695 | 9.422 | 10.059 |
3 | LLPTA | 8.476 | 8.832 | 8.654 |
4 | CPI | 8.070 | 6.506 | 7.288 |
5 | LIQR | 7.279 | 7.219 | 7.249 |
6 | DFR | 7.184 | 6.770 | 6.977 |
7 | ROA | 6.059 | 7.679 | 6.869 |
8 | FATA | 6.242 | 6.898 | 6.570 |
9 | GDP | 6.910 | 5.572 | 6.241 |
10 | LEAD | 6.708 | 5.160 | 5.934 |
11 | GVTTA | 5.916 | 5.909 | 5.912 |
12 | ECB | 3.866 | 3.209 | 3.538 |
13 | GVT2Y | 2.583 | 2.347 | 2.465 |
14 | ON | 2.436 | 2.463 | 2.450 |
15 | GVT9M | 2.171 | 2.373 | 2.272 |
16 | OIL | 1.382 | 1.285 | 1.334 |
17 | LENDINGR | 0.502 | 1.590 | 1.046 |
18 | FFR | 1.006 | 1.064 | 1.035 |
19 | REER | 0.897 | 1.068 | 0.982 |