ML//time series//EWMA

An exponentially weighted moving average (EWMA) is a running average in which recent samples count for more and old ones decay geometrically. Each new value bumps it up, it decays between values, and one tunable constant sets the length of its memory.

An exponentially weighted moving average (EWMA) is a running average in which recent samples count for more and old ones decay geometrically. Each new value bumps it up, it decays between values, and one tunable constant sets the length of its memory.

It is a cheap, causal feature for how active a stream is right now, and an observable stand-in for Hawkes intensity. In a dual-clock feature set it serves as the fast clock (the recent density), paired with a slower accumulator.