PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
January 1, 19984,238 citationsOpen Access

Approximate nearest neighbors

PIPiotr IndykRMRajeev Motwani

Key Points

Key points are not available for this paper at this time.

Abstract

The nearest neighbor problem is the follolving: Given a set of n points P = (PI, . . . ,p,} in some metric space X, preprocess P so as to efficiently answer queries which require finding bhe point in P closest to a query point q E X. We focus on the particularly interesting case of the d-dimensional Euclidean space where X = Wd under some Zp norm. Despite decades of effort, t,he current solutions are far from saabisfactory; in fact, for large d, in theory or in practice, they provide litt,le improvement over the brute-force algorithm which compares the query point to each data point. Of late, t,here has been some interest in the approximate newest neighbors problem, which is: Find a point p E P that is an c-approximate nearest neighbor of the query q in t,hat for all p' E P, d(p, q) 1); and, (b) query time polynomial in log-n and d, and only a mildly exponential preprocessing cost* O(n) x 0(1/~)~. Furt.her, applying a classical geometric lemma on random projections (for which we give a simpler proof), we obtain t.he first known algorithm with polynomial preprocessing and query t.ime polynomial in d and log n. Unfortunately, for small E, the latter is a purely theoretical result since bhe e?rponent depends on l/e. Experimental resuits indicate that our tit algori here, we give applications to information ret,rieval, pattern recognition, dynamic closest-pairs, and fast clustering algorithms.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Indyk et al. (1998) studied this question.

synapsesocial.com/papers/69d7d60f3b601d7be3ae3233https://doi.org/10.1145/276698.276876
Ask AI
Helpful
Bookmark
Share
View Full Paper