Research Article
A Pruning Neural Network Model in Credit Classification Analysis
Table 13
Extraction rate and accuracy rate comparison between PNN and other feature selection methods of the Australian credit dataset.
| Methods | Numbers of variables before feature selection | Numbers of variables after feature selection | Extraction rate (%) | Accuracy rate (%) |
| -test [54] | 14 | 12 | 14.3 | 89.27 | Stepwise [54] | 14 | 7 | 50 | 84.74 | Related matrix [54] | 14 | 12 | 14.3 | 89.31 | Factor analysis [54] | 14 | 9 | 35.7 | 86.08 | PCA [54] | 14 | 9 | 35.7 | 89.93 | PNN | 14 | 4 | 71.4 | 85.80 |
|
|