ML//time series//Hawkes process

A Hawkes process is a self-exciting point process: each event temporarily raises the probability of the next, so events arrive in bursts.

A Hawkes process is a self-exciting point process: each event temporarily raises the probability of the next, so events arrive in bursts.

The intensity jumps after every event and decays back down, so clusters emerge without any external trigger.

It models earthquake aftershocks, financial order flow and neural event streams.

The observable proxy for its intensity is often an EWMA of the event stream.