Hey @manyangled have you evaluated https://github.com/stanford-futuredata/momentsketch … vs t-digest?
-
-
Replying to @avibryant
Not yet! My expectation is that the moment based approach is lower fidelity, but it is going to be smaller and I'd expect it to be faster as well. Coincidentally I've been working on a faster (mutable) java t-digest, using a Fenwick Tree for maintaining cumulative masses.
1 reply 0 retweets 0 likes
Replying to @manyangled
It seems like it would cover the t-digest + monotonic spline fitting use case (since it's also doing an optimization trying to fit to some analytical distribution, though I'm unclear on what the space of dists is it looks at)
1:06 PM - 7 Nov 2018
0 replies
0 retweets
1 like
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.