I'm trying to get familiar with MinHash function, and there are a few things a do not understand:
- Why to use multiple hash function version over single hash function alternative? MinHash wiki says - both implementations of MinHash gives the same expected error. But in the same time single hash function is cheaper.
- Which hash functions I have to use in both versions of MinHash? Is it acceptable to use "cheaper" hash functions in multi-hash-function version?
- Are there any ways to tune accuracy/speed besides increasing amount of hash functions?