A nearest-neighbor approach to estimating divergence between continuous random vectors

被引：44
|
作者：
Wang, Qing [1 ]
Kulkarni, Sanjeev R. [1 ]
Verdu, Sergio [1 ]
机构：
[1] Princeton Univ, Dept Elect Engn, Princeton, NJ 08544 USA
来源：
2006 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY, VOLS 1-6, PROCEEDINGS | 2006年
基金：
美国国家科学基金会;
关键词：
D O I：
10.1109/ISIT.2006.261842
中图分类号：
TN [电子技术、通信技术];
学科分类号：
0809 ;
摘要：
A method for divergence estimation between multidimensional distributions based on nearest neighbor distances is proposed. Given i.i.d. samples, both the bias and the variance of this estimator are proven to vanish as sample sizes go to infinity. In experiments on high-dimensional data, the nearest neighbor approach generally exhibits faster convergence compared to previous algorithms based on partitioning.
引用
收藏
页码：242 / +
页数：3
相关论文
共 50 条
12345