CMinHash
A MinHash algorithm variant based on a cyclic permutation design. It generates multiple correlated hash functions via a circulant matrix, reducing variance while also reducing storage space.
Timeline (last 90 days)
CMinHash is a MinHash algorithm designed through circulant permutation, which can reduce both variance and storage space.
CMinHash requires only one random seed to generate multiple related but approximately independent hash functions.
The new analytical framework can cover CMinHash itself and can also handle more structurally complex estimators such as Minner, thereby demonstrating its generality.
All Facts (3)
CMinHash is a MinHash algorithm designed through circulant permutation, which can reduce both variance and storage space.
50%UnverifiedCMinHash requires only one random seed to generate multiple related but approximately independent hash functions.
50%UnverifiedThe new analytical framework can cover CMinHash itself and can also handle more structurally complex estimators such as Minner, thereby demonstrating its generality.
50%