Table 4.
Random forest hyperparameter tuning for each prediction task.
Target variable | Number of Trees |
Terminal Node Size |
Sampled Variables |
Validation AUC |
|
---|---|---|---|---|---|
TBI | 100 | 4 | 22 | 0.906 | |
200 | 1 | 18 | 0.9068 | ||
300 | 1 | 22 | 0.908 | ||
400 | 1 | 15 | 0.9062 | ||
Cause of Injury | Falls | 100 | 2 | 22 | 0.8954 |
200 | 1 | 22 | 0.8975 | ||
300 | 1 | 22 | 0.8987 | ||
400 | 1 | 22 | 0.8988 | ||
500 | 1 | 18 | 0.8988 | ||
Struck by/against | 100 | 1 | 6 | 0.892 | |
200 | 1 | 6 | 0.8969 | ||
300 | 2 | 5 | 0.8939 | ||
Cyclist Collision | 100 | 1 | 18 | 0.856 | |
200 | 1 | 18 | 0.8589 | ||
300 | 1 | 18 | 0.86 | ||
400 | 2 | 22 | 0.8664 | ||
500 | 2 | 12 | 0.8629 | ||
Motor Vehicle Collision | 100 | 4 | 22 | 0.8944 | |
200 | 2 | 22 | 0.8971 | ||
300 | 2 | 18 | 0.8981 | ||
400 | 2 | 23 | 0.8985 | ||
500 | 2 | 18 | 0.8989 | ||
600 | 1 | 18 | 0.8997 | ||
700 | 1 | 18 | 0.8996 |