A New Ensemble Diversity Measure Applied to Thinning Ensembles
A New Ensemble Diversity Measure Applied to Thinning Ensembles
复制标题
DOI：
10.1007/3-540-44938-8_31
复制
发表时间：
2003-06
期刊：
影响因子：
--
通讯作者：
Robert E. Banfield;L. Hall;K. Bowyer;W. Kegelmeyer
中科院分区：
文献类型：
--
作者：
Robert E. Banfield;L. Hall;K. Bowyer;W. Kegelmeyer
我们介绍了一种描述分类器集合多样性的新方法--正确率多样性度量，并将其与现有方法进行了比较。然后，我们介绍了两种基于多样性计算从集成中移除分类器的新方法。来自加州大学欧文分校知识库的12个数据集的经验结果表明，多样性通常是通过我们的测量来建模的，并且可以在不损失精度的情况下将集合变得更小。
We introduce a new way of describing the diversity of an ensemble of classifiers, the Percentage Correct Diversity Measure, and compare it against existing methods. We then introduce two new methods for removing classifiers from an ensemble based on diversity calculations. Empirical results for twelve datasets from the UC Irvine repository show that diversity is generally modeled by our measure and ensembles can be made smaller without loss in accuracy.