Approximate Nearest Neighbor Search (ANNS) Algorithm
Jump to navigation
Jump to search
A Approximate Nearest Neighbor Search (ANNS) Algorithm is a nearest neighbor search algorithm that efficiently finds the nearest neighbors to a query point in a dataset, where the results are approximate rather than exact.
- Context:
- It can (typically) be used in large-scale, high-dimensional datasets where exact nearest neighbor search is computationally expensive.
- It can (typically) prioritize speed and resource efficiency over absolute accuracy in the search results.
- It can (typically) be evaluated based on trade-offs between accuracy, query time, and memory usage.
- It can (often) utilize algorithms like Locality-Sensitive Hashing (LSH), Product Quantization (PQ), or tree-based methods like KD-trees and Ball trees.
- ...
- Example(s):
- a kNN Search Algorithm.
- ...
- Counter-Example(s):
- See: Nearest Neighbor Search, Locality-Sensitive Hashing, Product Quantization, KD-tree, Ball Tree, High-Dimensional Data.