k Approximate Nearest Neighbors Search (Nearest Neighbor Search)

From Algorithm Wiki
Revision as of 13:04, 15 February 2023 by Admin (talk | contribs)
Jump to navigation Jump to search

Description

Within a dataset of $n$ points, find approximately the $k$ closest points to a specified point.

Related Problems

Generalizations: k Nearest Neighbors Search

Parameters

n: number of points in dataset

k: number of neighbors to find

Table of Algorithms

Name Year Time Space Approximation Factor Model Reference
Hierarchical Navigable Small World (HNSW) 2018 $O(nlogn)$ $O(M)$ ? experimental results Deterministic Time & Space
Locality-sensitive hashing 2010 $O(nLkt)$ (pre-processing)

$O(L(kt+dnP_2^k))$ (query-time) || $O(nL)$ || c || Deterministic || Time & Space

Projected radial search 2013 $O(k)$ $O({1})$ ? Deterministic Time
Compression/Clustering (Vector Quantization) 1992 Varies by codebook structure Varies by codebook structure Deterministic