摘要
针对传统分类算法在处理非平衡数据集所出现的少数类分类准确率较低的问题,通过引入加权系数和样本分布函数给出了一种新的模糊规则权重的计算方法.该方法加强了类间的对比度和差异性,削弱了类内差距.将该权重方法与Chi et al规则生成算法和模糊分类推理模型结合形成新的分类算法,对具有不同非平衡度的UCI数据集进行Matlab对比研究,所得结果验证了该算法的可靠性与有效性.
For the problem that the traditional classification methods often tend to the majority class and lead a lower classification accuracy to the minority class in imbalancod data, a new calculation method of fuzzy role weights is proposed. This algorithm not only keeps the pattern matching degree within class in uniform distribution, but also enhances the contrast of inter-class. Then a classification algorithm is designed, which includes the new calculation method of fuzzy rule weights, Chi et al algorithm and fuzzy reasoning model. Finally numerical simulation about the imbalanced data of UCI data sets shows the reliability of the classification algorithm.
出处
《控制与决策》
EI
CSCD
北大核心
2012年第1期104-108,共5页
Control and Decision
基金
国家自然科学基金项目(60875032/F030504)
