Table 5 Random Forest feature importance scores for the prediction of high consumption enterprises
From: Explaining and predicting annual electricity demand of enterprises – a case study from Switzerland
Class | Class | Mean Decrease | Mean Decrease | |
|---|---|---|---|---|
low consumption | high consumption | Accuracy | Gini | |
brancheBFS | 1.92 | 21.85 | 8.61 | 167.74 |
log(area+1) | 10.44 | 33.86 | 21.35 | 309.43 |
number of facebook visits | 2.74 | 4.48 | 4.26 | 25.56 |
opening hours per week | 4.83 | 12.87 | 8.54 | 118.31 |
combined number of ratings | -4.99 | 10.27 | -1.451 | 30.36 |