Research Article
A Parallel Genetic Algorithm Based Feature Selection and Parameter Optimization for Support Vector Machine
Table 3
Comparisons between CGPGA-SVM and grid search.
| Dataset | CGPGA-SVM | Grid search | p values | Tenfold cross validation accuracy (%) | Number of selected features | Tenfold cross validation accuracy (%) | Number of selected features |
| Ionosphere | 98.85 ± 2.01 | 13.8 ± 3.46 | 93.16 ± 3.00 | 34.0 ± 0.00 | 0.005 | Breast cancer | 98.53 ± 1.16 | 2.60 ± 0.84 | 96.77 ± 1.81 | 10.0 ± 0.00 | 0.039 | Australia | 90.13 ± 2.39 | 5.10 ± 1.60 | 85.51 ± 4.46 | 14.0 ± 0.00 | 0.011 | Diabetes | 81.76 ± 3.37 | 3.90 ± 0.99 | 76.44 ± 3.74 | 8.00 ± 0.00 | 0.014 | Vehicle | 86.05 ± 3.54 | 10.3 ± 1.34 | 78.60 ± 3.29 | 18.0 ± 0.00 | 0.005 | Vowel | 99.29 ± 0.68 | 6.70 ± 0.48 | 98.88 ± 1.00 | 13.0 ± 0.00 | 0.037 | Car | 99.83 ± 0.39 | 6.00 ± 0.00 | 99.48 ± 0.51 | 6.00 ± 0.00 | 0.046 | Splice | 92.66 ± 1.43 | 26.7 ± 3.53 | 92.03 ± 1.52 | 60.0 ± 0.00 | 0.386 | DNA | 96.79 ± 1.31 | 87.1 ± 4.53 | 96.05 ± 1.09 | 180. ± 0.00 | 0.005 | WaveForm | 87.81 ± 1.60 | 21.1 ± 2.69 | 86.68 ± 1.47 | 40.0 ± 0.00 | 0.037 | Svmguide1 | 96.66 ± 0.76 | 2.40 ± 0.52 | 96.32 ± 0.71 | 4.00 ± 0.00 | 0.594 | Mushrooms | 100.0 ± 0.00 | 42.5 ± 3.54 | 99.98 ± 0.04 | 112. ± 0.00 | 0.317 |
|
|
indicates significance at 0.005 level.
|