On the computation of Shannon Entropy from Counting Bloom Filters

TL;DR

This paper presents a method to estimate discrete Shannon entropy using counting Bloom filters, effective when collision probability remains low, offering a practical approach for entropy computation in data structures.

Contribution

It introduces a novel method for computing the naive plugin entropy estimator directly from counting Bloom filters, which was not previously established.

Findings

Works reasonably with low collision probability

Provides a practical entropy estimation technique

Applicable to data structures with counting Bloom filters

Abstract

In this short note a method for computing the naive plugin estimator of discrete entropy from a counting Bloom filter will be presented. The method does work reasonably as long as the collision probability in the bloom filter is kept low.