Title :
A Nearest-Neighbor Approach to Estimating Divergence between Continuous Random Vectors
Author :
Wang, Qing ; Kulkarni, Sanjeev R. ; Verdu, Sergio
Author_Institution :
Dept. of Electr. Eng., Princeton Univ., NJ
Abstract :
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
Keywords :
statistical distributions; continuous random vectors; divergence estimation; multidimensional distributions; nearest-neighbor distances approach; Convergence; Entropy; Frequency estimation; H infinity control; Multidimensional systems; Nearest neighbor searches; Neural networks; Partitioning algorithms; Probability distribution; Random variables;
Conference_Titel :
Information Theory, 2006 IEEE International Symposium on
Conference_Location :
Seattle, WA
Print_ISBN :
1-4244-0505-X
Electronic_ISBN :
1-4244-0504-1
DOI :
10.1109/ISIT.2006.261842
