ABC-Boost：用于多类分类的自适应基类升压][ABC-Boost: Adaptive Base Class Boost for Multi-Class Classiﬁcation]

ABC-Boost：用于多类分类的自适应基类升压 ABC-Boost: Adaptive Base Class Boost for Multi-Class Classiﬁcation


课程网址:	http://videolectures.net/icml09_li_abcb/
主讲教师:	Ping Li
开课单位:	康奈尔大学
开课时间:	2009-08-26
课程语种:	英语
中文简介:	我们提出ABC Boost（自适应基类升压）用于多类分类，并提出ABC MART，ABC Boost的实现。原始的MART（多重加法回归树）算法在某些行业应用（例如，Web搜索）中很流行。对于二元分类，ABC MART恢复了MART。对于多类分类，ABC MART改进了MART，在几个公共数据集上进行了评估。
课程简介:	We propose ABC-Boost (Adaptive Base Class Boost) for multi-class classiﬁcation and present ABC-MART, an implementation of ABC-Boost. The original MART (Multiple Additive Regression Trees) algorithm has been popular in certain industry applications (e.g., Web search). For binary classiﬁcation, ABC-MART recovers MART. For multi-class classiﬁcation, ABC-MART improves MART, as evaluated on several public data sets.
关键词:	多类分类; Web搜索; 公共数据集
课程来源:	视频讲座网
最后编审:	2019-04-23：lxf
阅读次数:	229

境外开放课程导航