The Analysis of a Probabilistic Approach to Nearest Neighbor Searching
WADS '01 Proceedings of the 7th International Workshop on Algorithms and Data Structures
An Empirical Study of a New Approach to Nearest Neighbor Searching
ALENEX '01 Revised Papers from the Third International Workshop on Algorithm Engineering and Experimentation
Buoy indexing of metric feature spaces for fast approximate image queries
Proceedings of the sixth Eurographics workshop on Multimedia 2001
Hi-index | 0.00 |
We investigate the problem of approximate similarity (nearest neighbor) search in high-dimensional metric spaces, and describe how the distance distribution of the query object can be exploited so as to provide probabilistic guarantees on the quality of the result. This leads to a new paradigm for similarity search, called PAC-NN (probably approximately correct nearest neighbor) queries, aiming to break the "dimensionality curse". PAC-NN queries return, with probability at least \math a \math-approximate NN -- an object whose distance from the query q is less than \math times the distance between q and its NN. Analytical and experimental results obtained for sequential and index-based algorithms show that PAC-NN queries can be efficiently processed even on very high-dimensional spaces and that control can be exerted in order to tradeoff the accuracy of the result and the cost.