Tree ensemble construction using a GRASP-based heuristic and annealed randomness.

Information Fusion(2014)

引用 10|浏览13
暂无评分
摘要
•Two new tree ensemble construction methods are presented.•The methods use the same solution generation strategy that is used by GRASP.•One aims to increase the diversity while minimizing the effect on accuracy.•The randomness is high at the root in order to encourage diversity.•The randomness is low at leaves to reduce loss of accuracy.
更多
查看译文
关键词
GRASP metahuristic,Decision trees,Classifier ensembles,Boosting,Random Forest
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要