Two improved attribute weighting schemes for value difference metric

Two improved attribute weighting schemes for value difference metric
复制标题

两种改进的价值差异度量属性加权方案

DOI:
10.1007/s10115-018-1229-3
复制
发表时间:
2018-06
影响因子:
2.7
通讯作者:
Li Chaoqun
Li Chaoqun
中科院分区:
计算机科学4区
文献类型:
--
作者:
Jiang Liangxiao;Li Chaoqun

文献摘要

参考文献

被引文献

相似文献

由于其简单性、效率和功效,价值差异度量(VDM)在面对更复杂的新手时继续表现良好,因此仍然引起了距离度量学习社区的极大兴趣。在通过削弱属性独立性假设来改进 VDM 的众多方法中,属性加权受到的关注较少(只有两种属性加权方案),但表现出了显着的类别概率估计性能。现有的两种属性加权方案,一种是非对称的,另一种是对称的。在本文中,我们提出了两项​​简单的改进来设置与 VDM 一起使用的属性权重。一种是非对称Kullback-Leibler散度加权值差值度量(KLD-VDM),另一种是对称增益比加权值差值度量(GR-VDM)。我们对大量数据集进行了广泛的评估,发现 KLD-VDM 和 GR-VDM 在负条件对数似然和相对平方根误差方面显着优于两种现有的属性加权方案,同时保持了 VDM 所特有的计算简单性和鲁棒性。
Due to its simplicity, efficiency and efficacy, value difference metric (VDM) has continued to perform well against more sophisticated newcomers and thus has remained of great interest to the distance metric learning community. Of numerous approaches to improving VDM by weakening its attribute independence assumption, attribute weighting has received less attention (only two attribute weighting schemes) but demonstrated remarkable class probability estimation performance. Among two existing attribute weighting schemes, one is non-symmetric and the other is symmetric. In this paper, we propose two simple improvements for setting attribute weights for use with VDM. One is the non-symmetric Kullback–Leibler divergence weighted value difference metric (KLD-VDM) and the other is the symmetric gain ratio weighted value difference metric (GR-VDM). We performed extensive evaluations on a large number of datasets and found that KLD-VDM and GR-VDM significantly outperform two existing attribute weighting schemes in terms of the negative conditional log likelihood and root relative squared error, yet at the same time maintain the computational simplicity and robustness that characterize VDM.
局部值差异度量
DOI: 10.1016/j.patrec.2014.06.014
发表时间: 2014-11-01
影响因子: 5.1
作者:
Li, Chaoqun;Jiang, Liangxiao;Li, Hongwei
通讯作者: Li, Hongwei
DOI: --
发表时间: 2014
期刊: Journal of management science
影响因子: --
作者:
อนิรุธ สืบสิงห์
通讯作者: อนิรุธ สืบสิงห์
DOI: 10.1109/cvpr.2000.855863
发表时间: 2000
期刊: Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662)
影响因子: --
作者:
C. Domeniconi;D. Gunopulos;Jing Peng
通讯作者: C. Domeniconi;D. Gunopulos;Jing Peng
分类并不总是简单:学习 SuperParent 进行类概率估计
DOI: --
发表时间: 2015
影响因子: 8.5
作者:
Qiu, chen;Jiang, Liangxiao;Li, Chaoqun
通讯作者: Li, Chaoqun
DOI: --
发表时间: 1992-10
期刊: --
影响因子: --
作者:
J. R. Quinlan
通讯作者: J. R. Quinlan