@akIntelligence Hey AK, have you guys tried HLL + minhash to get intersections (http://tech.adroll.com/media/hllminhash.pdf …)? Looks interesting
@ZacharyTong @akIntelligence if you're spending the memory on a minhash anyway, I wonder how much benefit you get from the additional HLL?
-
-
@avibryant@akIntelligence I assumed it was one minhash per HLL? Since minhash only does set-similarity, HLL is still useful for cardinality -
@ZacharyTong@akintelligence right, but you can do cardinality estimation with minhash too (it's just less space-efficient than HLL). -
@avibryant@akintelligence Ah interesting, didn't realize you could get cardinality with just minhash. Only have passing experience with it -
@ZacharyTong@akIntelligence BTW what they use in that paper is not what I would call minhash, but rather KMV: http://blog.aggregateknowledge.com/2012/07/09/sketch-of-the-day-k-minimum-values/ … -
@avibryant@ZacharyTong@akIntelligence (shameless plug:@timonk did some great work on HLL intersections: http://blog.aggregateknowledge.com/2012/12/17/hll-intersections-2/ …)
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.