We present Classy Ensemble, a novel ensemble-generation algorithm for classification tasks, which aggregates models through a weighted combination of per-class accuracy. Tested over 153 machine learning datasets we demonstrate that Classy Ensemble outperforms two other well-known aggregation algorithms -- order-based pruning and clustering-based pruning -- as well as the recently introduced lexigarden ensemble generator. We also show preliminary results for deep networks.

翻译：我们提出了Classy Ensemble，一种面向分类任务的新颖集成生成算法，该算法通过基于各类别准确率的加权组合来聚合模型。通过对153个机器学习数据集的测试，我们证明Classy Ensemble的性能优于其他两种著名的聚合算法——基于排序的剪枝和基于聚类的剪枝——以及近期提出的lexigarden集成生成器。我们还展示了针对深度网络的初步结果。