Research Article
Density-Based Geometric One-Class Classifier Combined with Genetic Algorithm
Table 3
Result of pretest for choosing kernel function.
| | Kernel type | Dataset | | Iris (Virginica) | Biomed | Liver | Breast | Vehicle | Abalone | Satellite | | Average number of generations until convergence (standard deviation) |
| | Uniform kernel | 6.75 (1.43) | 7.80 (1.65) | 7.25 (1.05) | 7.65 (1.58) | 11.35 (1.92) | 13.45 (2.54) | 17.55 (2.46) | | Triangular kernel | 6.95 (1.52) | 7.95 (1.72) | 7.20 (0.98) | 7.40 (1.46) | 10.85 (2.09) | 12.95 (2.83) | 18.00 (2.33) | | Gaussian kernel | 6.90 (1.33) | 7.55 (1.24) | 7.10 (0.99) | 7.45 (1.64) | 8.05 (1.42) | 9.05 (1.52) | 11.05 (1.88) | | Cosine kernel | 6.80 (1.67) | 7.70 (1.31) | 7.15 (1.06) | 7.55 (1.54) | 8.00 (1.35) | 8.95 (1.44) | 10.95 (2.01) |
|
|
