ML//time series//point process

A point process is a stochastic model of the random timing of discrete events on a line (or in space), used for data that arrive as events rather than as values sampled on a fixed clock. What is modelled is the intensity (the instantaneous rate), not the events directly.

A point process is a stochastic model of the random timing of discrete events on a line (or in space), used for data that arrive as events rather than as values sampled on a fixed clock. What is modelled is the intensity (the instantaneous rate), not the events directly.

The baseline is the Poisson process, where events arrive independently at a constant rate. A Hawkes process adds self-excitation, and other variants add inhibition or external drivers.