Research Article
Feature Selection Method Based on Artificial Bee Colony Algorithm and Support Vector Machines for Medical Datasets Classification
Table 10
Classification accuracies obtained by our method and other classifiers for diabetes dataset.
| Author (year) | Method | Classification accuracy (%) |
| Şahan et al. (2005) [14] | AWAIS (10-fold CV) | 75.87 | Polat and Güneş (2007) [15] | Combining PCA and ANFIS | 89.47 |
Polat et al. (2008) [16] | LS-SVM (10-fold CV) | 78.21 | GDA-LS-SVM (10-fold CV) | 82.05 | Kahramanli and Allahverdi (2008) [17] | Hybrid system (ANN and FNN) | 84.2 | Patil et al. (2010) [18] | Hybrid prediction model (HPM ) with reduced dataset | 92.38 | Isa and Mamat (2011) [19] | Clustered-HMLP | 80.59 | Aibinu et al. (2011) [20] | AR1 + NN (3-fold CV) | 81.28 | Our study | ABCFS + SVM (train: 75%-test: 25%) | 86.97 | ABCFS + SVM (10-fold CV) | 79.29 |
|
|