Research Article
Multiple Naïve Bayes Classifiers Ensemble for Traffic Incident Detection
Table 4
Experimental Results of NB, Five Rules and NBTree Based as Applied to the AYE Dataset (The performances are presented in the form average ± variance).
| Algorithm | DR | FAR | MTTD | CR | AUC | Kappa | MAE | RMSE | EC |
| Naïve Bayes classifier | 0.7112 ± 0.00348 | 0.0499 ± 8.16E − 04 | 1.394 ± 0.08869 | 0.9094 ± 0.00111 | 0.8663 ± 4.39E − 04 | 0.6247 ± 9.14E − 04 | 0.16769 ± 3.07E − 04 | 0.57895 ± 5.31E − 04 | 0.90854 ± 1.83E − 04 |
| Product rule ensemble | 0.6457 ± 0.00127 | 0.0557 ± 2.40E − 06 | 1.8685 ± 1.24139 | 0.8732 ± 1.64E − 06 | 0.8619 ± 4.85E − 07 | 0.7479 ± 3.58E − 05 | 0.16690 ± 1.04E − 06 | 0.57775 ± 3.11E − 06 | 0.90895 ± 3.68E − 07 |
| Sum rule ensemble | 0.7923 ± 5.57E − 05 | 0.0500 ± 6.73E − 07 | 1.4384 ± 0.63868 | 0.8774 ± 3.55E − 07 | 0.8667 ± 3.75E − 07 | 0.7439 ± 1.29E − 04 | 0.16739 ± 4.94E − 07 | 0.57861 ± 1.47E − 06 | 0.90866 ± 1.75E − 07 |
| Max rule ensemble | 0.7869 ± 1.79E − 04 | 0.0500 ± 2.21E − 06 | 1.4571 ± 0.72272 | 0.8771 ± 1.25E − 06 | 0.8663 ± 2.02E − 07 | 0.7511 ± 4.12E − 05 | 0.16803 ± 4.96E − 06 | 0.57969 ± 1.46E − 05 | 0.90828 ± 1.77E − 06 |
| Min Rule ensemble | 0.7960 ± 1.25E − 04 | 0.0498 ± 2.76E − 06 | 1.4646 ± 0.73539 | 0.8781 ± 1.10E − 06 | 0.8667 ± 1.14E − 06 | 0.6052 ± 1.86E − 06 | 0.16638 ± 3.25E − 06 | 0.57684 ± 9.83E − 06 | 0.90926 ± 1.15E − 06 |
| MV rule ensemble | 0.6246 ± 1.15E − 05 | 0.0572 ± 2.29E − 06 | 1.8358 ± 1.19124 | 0.8720 ± 3.84E − 07 | 0.7837 ± 1.18E − 06 | 0.7341 ± 5.31E − 05 | 0.16687 ± 2.27E − 06 | 0.57769 ± 6.87E − 06 | 0.90897 ± 8.03E − 07 |
| NBTree classifier | 0.7275 ± 3.25E − 4 | 0.0250 ± 1.85E − 5 | 1.2103 ± 0.10383 | 0.9031 ± 3.52E − 5 | 0.8512 ± 1.04E − 4 | 0.7085 ± 3.26E − 4 | 0.1711 ± 6.42E − 5 | 0.49189 ± 2.62E − 4 | 0.90553 ± 2.06E − 5 |
|
|