Top 4.7% forks on proxy.golang.org
proxy.golang.org : github.com/lytics/hll
This is a Go implementation of the HyperLogLog++ algorithm from "HyperLogLog in Practice: Algorithmic Engineering of a State of The Art Cardinality Estimation Algorithm" by Heule, Nunkesser and Hall of Google. This is a cardinality estimation algorithm: given a stream of input elements, it will estimate the number of unique items in the stream. The estimation error can be controlled by choosing how much memory to use. HyperLogLog++ improves on the basic HyperLogLog algorithm by using less space, improving accuracy, and correcting bias. This code is a translation of the pseudocode contained in Figures 6 and 7 of the Google paper. Not all algorithms are provided in the paper, but we've tried our best to be true to the authors' intent when writing the omitted algorithms. We're not trying to be creative, we're just implementing the algorithm described in the paper as directly as possible. The HyperLogLog++ paper is available at http://static.googleusercontent.com/media/research.google.com/en/us/pubs/archive/40671.pdf
Registry
-
Source
- Documentation
- JSON
purl: pkg:golang/github.com/lytics/hll
License: Apache-2.0
Latest release: about 7 years ago
First release: about 7 years ago
Namespace: github.com/lytics
Stars: 43 on GitHub
Forks: 6 on GitHub
See more repository details: repos.ecosyste.ms
Last synced: about 1 month ago