[GitHub] [datasketches-java] PavelVesely commented on issue #357: Performance comparison between DataSketches and DDSketch

[GitBox <gi...@apache.org>]

PavelVesely commented on issue #357:
URL: https://github.com/apache/datasketches-java/issues/357#issuecomment-889424997


   The fact that DDSketch paper cites the Agarwal et al. paper doesn't mean that DDSketch is based on the ideas from the Agarwal et al. paper. As far as I know, it is quite different.
   
   As mentioned above, the major difference between DDSketch and KLL/ReqSketch is indeed in the guarantee offered. For DDSketch, the guarantee is in the value space, that is, the *value* of the estimated quantile y returned by the sketch is by at most epsilon * x different from the true quantile x. On the other hand, KLL/ReqSketch offer guarantee in the rank space, that is, on quantile query q, they return an input item with rank close to q. The difference between these two guarantees is explained for instance in Section 2.1 in [this paper](https://arxiv.org/abs/2102.09299).
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@datasketches.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@datasketches.apache.org
For additional commands, e-mail: commits-help@datasketches.apache.org