Table 7. Performance results achieved by the models built with different machine learning algorithms.
| Metric | Machine learning algorithms | ||||||
|---|---|---|---|---|---|---|---|
| k-NN | Decision tree | Random forest | SVM | MLP | MLPE | ||
| Overall accuracy (%) | 98.6 | 95.1 | 98.6 | 98.3 | 98.3 | 98.4 | |
| Overall F1 score (%) | |||||||
| Training time (min) | 0.0 ± 0.0 | 0.1 ± 0.0 | 3.3 ± 0.4 | 0.5 ± 0.0 | 1.7 ± 0.1 | 1.5 ± 0.1 | |
| Prediction time | 1 frame (ms) | 349 ± 165 | 3 ± 7 | 67 ± 44 | 10 ± 10 | 5 ± 8 | 12 ± 9 |
| 1,800 frames (min) | 10.5 ± 4.9 | 0.1 ± 0.2 | 2.0 ± 1.3 | 0.3 ± 0.3 | 0.2 ± 0.2 | 0.4 ± 0.3 | |
The models’ performance results include the overall accuracy and overall F1 score, as well as the mean and standard deviation values for the training time when considering five runs, and the prediction time for a single frame and for 1,800 frames (≈1 min of data).