Research Article
A Differential Evolution-Oriented Pruning Neural Network Model for Bankruptcy Prediction
Table 12
Classification accuracies for Qualitative Bankruptcy dataset obtained by other hybrid classification methods in relative literatures.
| Author (year) | Method (train-to-test ratios) | Average accuracy (%) |
| Yi Tan et. al. (2016) [44] | Hybrid logistic regression-naive bayes (90%-10%)) | 99.64 | Nanxi Wang (2017) [45] | Neural network model with robust logistic regression (50%-50%) | 69.44 | Nanxi Wang (2017) [45] | Neural network model with inductive learning algorithm (50%-50%) | 89.7 | Nanxi Wang (2017) [45] | Neural network model with genetic algorithm (50%-50%) | 94 | Nanxi Wang (2017) [45] | Neural network model with neural networks without dropout (50%-50%) | 90.3 | Nanxi Wang (2017) [45] | Neural network model with SVM (50%-50%) | 98.67 | Nanxi Wang (2017) [45] | Neural network model with decision tree (50%-50%) | 99.33 | J.Uthayakumar et. al. (2018) [43] | Genetic ant colony algorithm (Not mentioned) | 91.32 | J.Uthayakumar et. al. (2018) [43] | Fitness-scaling chaotic Genetic ant colony algorithm (Not mentioned) | 92.14 | J.Uthayakumar et. al. (2018) [43] | Improved K-means clustering and fitness-scaling | 97.93 | chaotic genetic ant colony algorithm (Not mentioned) | Our Method (2019) | EPNN (50%-50%) | 99.57 | Our Method (2019) | EPNN (10 fold cross validation) | 99.68 |
|
|