| 用于分类的随机森林和Bagging分类树比较 |
| |
| 引用本文： | 马景义,谢邦昌. 用于分类的随机森林和Bagging分类树比较[J]. 统计与信息论坛, 2010, 25(10): 18-22 |
| |
| 作者姓名： | 马景义 谢邦昌 |
| |
| 作者单位： | 北京师范大学社会发展与公共政策学院;中央财经大学统计学院;辅仁大学统计资讯学系暨应用统计所; |
| |
| 基金项目： | 中央财经大学"121"人才工程青年博士发展基金，全国统计科学研究计划项目，教育部人文社会科学研究项目，中央财经大学学科建设基金 |
| |
| 摘 要： | 借助试验数据,从两种理论分析角度解释随机森林算法优于Bagging分类树算法的原因。将两种算法表述在两种不同的框架下,消除了这两种算法分析中的一些模糊之处。尤其在第二种分析框架下,更能清楚的看出,之所以随机森林算法优于Bagging分类树算法,是因为随机森林算法对应更小的偏差。
|
| 关 键 词： | 组合方法 随机森林 Bagging分类树 |
A Comparison on Random Forest and Bagging Classification Tree in Classification |
| |
| MA Jing-yi,XIE Bang-chang. A Comparison on Random Forest and Bagging Classification Tree in Classification[J]. Statistics & Information Tribune, 2010, 25(10): 18-22 |
| |
| Authors: | MA Jing-yi XIE Bang-chang |
| |
| Affiliation: | MA Jing-yi1,2,XIE Bang-chang2,3(1.School of Social Development and Public Policy,Beijing Normal University,Beijing 100875,China,2.School of Statistics,Central University of Finance and Economics,Beijing 100081,3.Dept.of Statistics and Information Science,Fu Jen Catholic University,Taipei 24205,China) |
| |
| Abstract: | The purpose of this paper is to explain why random forest is more competitive than Bagging classification tree in classification from two theoretical views by experiment data.Our analysis in two theoretical views may dissipate misunderstandings existing in the two ensemble methods,and the second analysis suggested by us shows random forest is more successful than bagging in bias reduction and may more evidently explain the reason that random forest is prior to Bagging classification tree than the first one. |
| |
| Keywords: | ensemble methods random forest bagging classification tree |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
