Module streaming 

Online (streaming) estimators that run in bounded memory.

Each estimator consumes one value at a time through update and reports its current estimate through an accessor, holding only a fixed-size summary of the stream seen so far. State size is a compile-time constant — independent of how many values have been consumed — so an arbitrarily long stream is processed without memory growth.

• RunningMoments — Welford’s online mean and variance.
• P2Quantile — the P² streaming-quantile estimator (Jain & Chlamtac 1985).

Structs

P2Quantile

Online estimator of a single p-quantile using the P² algorithm.

RunningMoments

Welford’s online accumulator for the running mean and variance of a stream.