Research Article
A Hybrid Feature Selection Method Based on Rough Conditional Mutual Information and Naive Bayesian Classifier
Table 4
Number and accuracy of selected features based on naive Bayes.
| Number | Unselect | RCMI | Filter wrapper | Accuracy | Feature number | Accuracy | Feature number | Accuracy | Feature number |
| 1 | 75.00% | 279 | 77.43% | 16 | 78.76% | 9 | 2 | 84.52% | 19 | 85.13% | 8 | 85.81% | 3 | 3 | 90.60% | 34 | 94.87% | 6 | 94.87% | 6 | 4 | 91.52% | 19 | 93.07% | 3 | 94.33% | 3 | 5 | 85.58% | 60 | 86.06% | 15 | 86.54% | 6 | 6 | 92.09% | 35 | 91.22% | 24 | 92.24% | 10 | 7 | 90.11% | 16 | 95.63% | 2 | 95.63% | 1 | 8 | 95.78% | 16 | 95.43% | 4 | 95.43% | 4 | 9 | 98.31% | 12 | 98.88% | 5 | 96.07% | 3 | 10 | 95.05% | 16 | 95.05% | 10 | 95.05% | 8 |
| Average | 89.86% | 50.6 | 91.28% | 9.3 | 91.47% | 5.3 |
|
|