Table 2 Summary of results of models predicting the traffic status after 6 h on our study’s street.
AdaBoost | Neural net. | Gradient boost. | Nearest neighbors | Support vect. mach. | |
---|---|---|---|---|---|
Accuracy score | 0.9805 | 0.9840 | 0.9826 | 0.9840 | 0.9826 |
F score (low traffic) | 0.9927 | 0.9927 | 0.9927 | 0.9943 | 0.9927 |
F score (mild traffic) | 0.9525 | 0.9584 | 0.9564 | 0.9608 | 0.9564 |
F score (high traffic) | 0.9825 | 0.9873 | 0.9855 | 0.9855 | 0.9855 |
Jaccard score (low traffic) | 0.9856 | 0.9856 | 0.9856 | 0.9886 | 0.9856 |
Jaccard score (mild traffic) | 0.9093 | 0.9203 | 0.9165 | 0.9245 | 0.9165 |
Jaccard score (high traffic) | 0.9657 | 0.9750 | 0.9714 | 0.9714 | 0.9714 |
First important predictor | t12 | t12 & t16 | t0 | t0 | t12 |
Second important predictor | t0 | t11 | t11 & t5 | t18 | t18 |