Research Article
Global Optimization Ensemble Model for Classification Methods
Table 5
Data set and suitable classifiers.
| Dataset | Classifier | Capabilities |
| All datasets | K-NN | Polynomial, numerical, binomial attributes, and labels can handle missing values |
| All datasets | Decision tree | Polynomial, numerical, and binomial attributes cannot handle numeric labels and can handle missing values | Heart, wine, and educational and sonar dataset | Rule induction |
| Cancer, heart, adult income dataset | ID3 | Can only handle binomial and polynomial labels and attributes and cannot handle missing values | All datasets | W-AODE | All datasets | W-Prism |
| Educational progress and sonar and adult income dataset | Random forest | Polynomial, numerical, and binomial attributes cannot handle numeric labels and cannot handle missing values | All datasets | W-PART | All datasets | W-J48 |
| Sonar, diabetes, cancer, andadult income dataset | Logistic regression | Numerical attributes and binomial labels cannot handle missing values |
|
|