Research Article
Identification of the Causative Disease of Intermittent Claudication through Walking Motion Analysis: Feature Analysis and Differentiation
Table 3
Classification results at every classifier.
(a) LDA (linear discriminant analysis) |
| Class of test data | Normal | PAD | L5 | L4 | All |
| Success | 10 | 8 | 10 | 2 | 30 | Failure | 3 | 2 | 3 | 8 | 16 |
| Total | 13 | 10 | 13 | 10 | 46 | Accuracy | 0.77 | 0.8 | 0.77 | 0.2 | 0.65 |
|
|
(b) Decision tree |
| Class of test data | Normal | PAD | L5 | L4 | All |
| Success | 7 | 5 | 5 | 1 | 18 | Failure | 6 | 5 | 8 | 9 | 28 |
| Total | 13 | 10 | 13 | 10 | 46 | Accuracy | 0.54 | 0.5 | 0.38 | 0.1 | 0.39 |
|
|
(c) One-versus-one SVM (OVO) |
| Class of test data | Normal | PAD | L5 | L4 | All |
| Success | 12 | 8 | 9 | 3 | 32 | Failure | 1 | 2 | 4 | 7 | 14 |
| Total | 13 | 10 | 13 | 10 | 46 | Accuracy | 0.92 | 0.8 | 0.69 | 0.3 | 0.70 |
|
|
(d) One versus-the-rest SVM (OVR) |
| Class of test data | Normal | PAD | L5 | L4 | All |
| Success | 13 | 8 | 9 | 1 | 31 | Failure | 0 | 2 | 4 | 9 | 15 |
| Total | 13 | 10 | 13 | 10 | 46 | Accuracy | 1 | 0.8 | 0.69 | 0.1 | 0.67 |
|
|
(e) Presented SVM-based methodology |
| Class of test data | Normal | PAD | L5 | L4 | All |
| Success | 11 | 9 | 11 | 7 | 38 | Failure | 2 | 1 | 2 | 3 | 8 |
| Total | 13 | 10 | 13 | 10 | 46 | Accuracy | 0.85 | 0.9 | 0.85 | 0.7 | 0.83 |
|
|