Theory and Algorithm for Learning with Dissimilarity Functions
Theory and Algorithm for Learning with Dissimilarity Functions
复制标题
相异函数学习的理论和算法
DOI:
10.1162/neco.2008.08-06-805
复制
发表时间:
2009-05
影响因子:
2.9
通讯作者:
Sugiyama, Masashi
中科院分区:
文献类型:
--
作者:
Wang, Liwei;Hatano, Kohei;Feng Jufu;Yang, Cheng;Sugiyama, Masashi
We study the problem of classification when only a dissimilarity function between objects is accessible. That is, data samples are represented not by feature vectors but in terms of their pairwise dissimilarities. We establish sufficient conditions for dissimilarity functions to allow building accurate classifiers. The theory immediately suggests a learning paradigm: construct an ensemble of simple classifiers, each depending on a pair of examples; then find a convex combination of them to achieve a large margin. We next develop a practical algorithm referred to as dissimilarity-based boosting (DBoost) for learning with dissimilarity functions under theoretical guidance. Experiments on a variety of databases demonstrate that the DBoost algorithm is promising for several dissimilarity measures widely used in practice.
登录
查看更多内容
DOI:
10.1016/c2009-0-27872-x
发表时间:
1972
期刊:
影响因子:
--
作者:
H. Shimodaira;Iain Murray
通讯作者:
Iain Murray
DOI:
10.1002/0470854774.ch1
发表时间:
2006
期刊:
--
影响因子:
--
作者:
P. Pudil;P. Somol;M. Haindl
通讯作者:
P. Pudil;P. Somol;M. Haindl
DOI:
10.1007/978-3-540-30215-5_16
发表时间:
2004-10
期刊:
--
影响因子:
--
作者:
Maria-Florina Balcan;Avrim Blum;S. Vempala
通讯作者:
Maria-Florina Balcan;Avrim Blum;S. Vempala
影响因子:
7.5
作者:
Schapire, RE;Singer, Y
通讯作者:
Singer, Y
影响因子:
7.5
作者:
Maria-Florina Balcan;Avrim Blum;S. Vempala
通讯作者:
Maria-Florina Balcan;Avrim Blum;S. Vempala