6 ms·
Hyperloglog analyses generally assume access to the full data stream, and so are O(n) at a minimum. Perhaps by running hyperloglog on a sublinear sample of the
by qyph 2y ago
Hyperloglog analyses generally assume access to the full data stream, and so are O(n) at a minimum. Perhaps by running hyperloglog on a sublinear sample of the dataset you'd get an algorithm in this class.
- doormatt 2y agoThat makes sense, thanks for explaining!