MinHash
Definition
MinHash estimates the Jaccard similarity of large sets via compact probabilistic signatures.
A probabilistic method for compact estimation of the Jaccard similarity of large sets.
Simply explained
Like comparing short samples of two large collections instead of checking every item individually.
Why it matters in practice
MinHash estimates set similarity with compact signatures. This allows more efficient narrowing of candidates for near-duplicate documents.
Example from business
Near-duplicate web pages are grouped using shingles and MinHash signatures.
Limits and pitfalls
- Estimation errors, choice of shingles, and unsuitability for other similarity measures are overlooked.
Origin & history
Broder introduced MinHash in the 1990s for detecting similar documents.
Use in mid-sized companies
- Possible application: A search team defines shingles and signature length to detect similar documents. It checks samples of the hits for misclassifications before merging documents.
Frequently asked questions
What does MinHash mean?
A probabilistic method for compact estimation of the Jaccard similarity of large sets.
Which mistakes should be avoided?
Estimation errors, choice of shingles, and unsuitability for other similarity measures are overlooked.
Sources
- dl.acm.org
- Ergänzender Primärbeleg · Ergänzender Primärbeleg
- Ergänzender Primärbeleg · Ergänzender Primärbeleg
Source checked on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results. Supplementary references checked on 12.10.2026.
Editorially reviewed on: