libtwiddle
libtwiddle copied to clipboard
fsaintjacques
bit & sketches data structures
- original estimator with elementary alpha correction - [google] proposed interpolation - [lookup] table - best fitting [polynomial] [google] Stefan Heule, Marc Nunkesser, and Alexander Hall. Hyperloglog in practice: Algorithmic...
This structure should be somewhat easy to implement since it mixes concept from hyperloglog and bloomfilter, both of which are already implemented. https://en.wikipedia.org/wiki/Count-min_sketch http://dimacs.rutgers.edu/~graham/pubs/papers/cmencyc.pdf
See basic java implementation https://github.com/metamx/extendedset.git and http://ricerca.mat.uniroma3.it/users/colanton/docs/concise.pdf
Current data structures don't offer a way to dump/load to memory/file/byte array. This will be crucial for production environment. This might force decoupling an `_init()` function from the current `_new()`.