基于集成学习的覆盖算法

Cover Algorithm Based on Ensemble Learning

介绍了传统的领域覆盖算法和交叉算法,并分析它们各自存在的缺点以及造成这些缺点的原因。针对传统的领域覆盖算法存在的泛化能力不足以及交叉覆盖算法存在的正确率不高的问题,提出了一种新的改进算法——基于集成学习的覆盖算法(CABE)。CABE是利用集成学习来整合交叉覆盖算法和领域覆盖算法,是通过对领域覆盖算法中的拒识样本的处理来提升算法的精度。使用UCI数据集进行实验,实验结果表明,改进的算法提高了算法分类的精度。

Introduces the traditional CA（ Cover Algorithm） and the CCA （Cross Cover Algorithm）, and analyses the shortcomings of them and the reasons. For low generalization ability of CA and low accuracy of CCA, propose an improved algorithm - CABE（Cover Algorithm Based on Ensemble）. CABE integrates CA and CCA through ensemble learning to improve performance through dealing with those samples that can＇t be classified. Finally the experimental results show that the CABE can improve the accuracy of classification.