Title :
Random Projection Trees for Vector Quantization
Author :
Dasgupta, Sanjoy ; Freund, Yoav
Author_Institution :
Dept. of Comput. Sci. & Eng., Univ. of California, La Jolla, CA
fDate :
7/1/2009 12:00:00 AM
Abstract :
A simple and computationally efficient scheme for tree-structured vector quantization is presented. Unlike previous methods, its quantization error depends only on the intrinsic dimension of the data distribution, rather than the apparent dimension of the space in which the data happen to lie.
Keywords :
trees (mathematics); vector quantisation; data distribution; quantization error; random projection trees; tree-structured vector quantization; Algorithm design and analysis; Computer science; Euclidean distance; Machine learning; Manifolds; Partitioning algorithms; Source coding; Statistical analysis; Statistics; Vector quantization; Computational complexity; manifolds; random projection; source coding; vector quantization;
Journal_Title :
Information Theory, IEEE Transactions on
DOI :
10.1109/TIT.2009.2021326
