Research Article
Hybrid Feature Selection Based Weighted Least Squares Twin Support Vector Machine Approach for Diagnosing Breast Cancer, Hepatitis, and Diabetes
Table 5
Comparison of predictive accuracies of proposed and other classifiers for Pima Diabetes.
| Reference | Approach | Predictive accuracy |
| This Study | HFS + WLSTSVM | 89.71% |
Polat et al. [48] | LSSVM | 78.21% | Temurtas et al. [36] | MLNN with LM | 82.37% | Kahramanli and Allahverdi [49] | Hybrid system | 84.2% |
Statlog [47] | Logdisc | 77.7% |
Jankowski [50] | IncNet | 77.6% |
Weka [47] | Logistic | 77.08% |
Statlog [47] | SMART | 76.8% | Weka [47] | Naïve Bayes | 76.04% | Weka [47] | SMO | 76.43% | Statlog [47] | BP | 75.2% |
Ster and Dobnikar [27] | ASI | 76.6% | Ster and Dobnikar [27] | MLP + BP | 76.4% | Ster and Dobnikar [27] | Fisher disc. analysis | 76.5% | Ster and Dobnikar [27] | QDA | 59.5% | Ster and Dobnikar [27] | KNN | 71.9% | Ster and Dobnikar [27] | CART | 72.8% |
Zarndt [47] | Bayes | 72.2 ± 6.9 | Zarndt [47] | C4.5 DT | 72.7 ± 6.6 | Zarndt [47] | ID3 | 71.7 ± 6.6 | Zarndt [47] | IB3 | 71.7 ± 5.0 | Zarndt [47] | OCN2 | 65.1 ± 1.1 |
|
|