Research Article
Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization
Table 1
Comparison of basic, resampling, and ADABOOST versions of SVM and MLP.
| Dataset | MLP | SVM | RMLP | RSVM | ADA-MLP | ADA-SVM |
| Cleveland | 79.2079 | 82.8383 | 93.72 | 85.0886 | 76.23 | 82.5083 | Statlog | 77.4074 | 84.07 | 86.6667 | 83.7037 | 77.777 | 84.07 | Spect | 79.4 | 81.65 | 88.38 | 88.764 | 79.4007 | 80.8989 | Spectf | 76.03 | 79.40 | 90.2622 | 82.397 | 76.03 | 77.9026 | Eric | 77.99 | 78.95 | 88.51 | 83.73 | 77.9904 | 77.9904 | WBC | 95.28 | 96.85 | 97.1388 | 96.5665 | 95.5651 | 96.7096 | Hepatitis | 81.94 | 85.16 | 90.3226 | 85.8065 | 78.7097 | 78.9097 | Thyroid | 96.28 | 89.77 | 98.1395 | 78.6047 | 97.2093 | 85.1163 | Parkinson | 91.28 | 86.15 | 96.4103 | 90.2564 | 92.3077 | 87.6923 | Pima Indian diabetics | 75.13 | 77.47 | 79.2969 | 76.0417 | 73.9583 | 77.3438 | BUPA liver | 71.59 | 70.14 | 68.1159 | 54.4928 | 71.3043 | 62.029 |
|
|
R: filter-based supervised instance resampling; ADA: ADABOOST.
|