摘要
In order to develop an objective and competitive credit scoring model, an immune network classification algorithm is proposed in this study. The algorithm employs information entropy minimization principle to transform the mixed input space into discretized space, and uses radial basis function to improve the system nonlinearity, and finally optimizes the immune classifier by pruning the cells with low fitness scores. The algorithm is used for Australian Credit Approval classification, the classification performance is evaluated by sensitivity, specificity and accuracy. The results show that our algorithm achieves better performance as compared with other algorithms, which means that our algorithm is a high-performance classifier and can provide valuable information for credit scoring to avoid making incorrect decisions.
- 出版日期2015
- 单位中南大学