Research Article
Supervised Expert System for Wearable MEMS Accelerometer-Based Fall Detector
Table 4
Performance by varying the kernel functions.
| ā | Sensitivity (%) | Specificity (%) | Relative support vectors | Relative execution time |
| Linear () | 98.6 | 52.3 | 0.51x | 0.2x | Linear () | 97.2 | 46.4 | 0.44x | 0.2x | Polynomial (, ) | 97.5 | 81.25 | 0.94x | 0.8x | Polynomial (, ) | 97.7 | 95.8 | 1x | 1x | GRBF (, ) | 98.1 | 83.1 | 1.1x | 1.3x | GRBF (, ) | 97.4 | 95.2 | 1.3x | 1.5x | Sigmoid (, , ) | 98.8 | 36 | 0.3x | 4x | Sigmoid (, , ) | 99.1 | 42 | 0.4x | 4.1x |
|
|