TABLE 2.
Ranking and feature importance score for different machine learning modelsa
| Features or independent variables | Adaptive Boosting Classifier | Decision Tree Classifier | Gaussian Naive Bayes | Logistic Regression | Multi-layer Perceptron Classifier | Random Forest Classifier | Stochastic Gradient Descent Classifier | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ranking | FIS | Ranking | FIS | Ranking | FIS | Ranking | FIS | Ranking | FIS | Ranking | FIS | Ranking | FIS | |
| gapA | 6 | 0.10 | 3 | 0.14 | 1 | 0.59 | 8 | 0.10 | 6 | 10.47 | 8 | 0.09 | 8 | 4.05 |
| ttr | 7 | 0.08 | 6 | 0.02 | 4 | 0.40 | 3 | 0.84 | 7 | 9.89 | 4 | 0.11 | 2 | 18.56 |
| usg | 3 | 0.14 | 4 | 0.13 | 2 | 0.46 | 4 | 0.74 | 5 | 10.72 | 7 | 0.09 | 5 | 10.11 |
| ttr/gapA | 1 | 0.20 | 2 | 0.21 | 5 | 0.19 | 1 | 0.94 | 1 | 11.81 | 3 | 0.13 | 3 | 14.58 |
| usg/gapA | 4 | 0.12 | 5 | 0.08 | 6 | 0.13 | 5 | 0.64 | 8 | 9.68 | 5 | 0.11 | 4 | 14.11 |
| Litter pH | 8 | 0.08 | 8 | 0.01 | 3 | 0.44 | 2 | 0.92 | 3 | 10.88 | 6 | 0.10 | 6 | 8.21 |
| Caked litter moisture | 2 | 0.16 | 1 | 0.39 | 7 | 0.02 | 6 | 0.54 | 4 | 10.81 | 1 | 0.24 | 1 | 23.13 |
| Friable litter moisture | 5 | 0.12 | 7 | 0.02 | 8 | 0.02 | 7 | 0.14 | 2 | 11.28 | 2 | 0.13 | 7 | 5.86 |
a
FIS, feature importance score.