A random process that represents occurrences as a collection of discrete points in a mathematical space (time, line, plane, or higher-dimensional domain), possibly with marks attached to points.

Principle

Characterization uses an intensity (first-moment measure), higher-order correlation measures, and properties such as complete randomness (Poisson), stationarity, or conditional intensity; independence and clustering properties distinguish common classes.

Demonstration

A temporal Poisson point process models call arrivals: events occur on the real line with independent increments and constant rate; a spatial Poisson process models tree locations in a uniform forest, while a Cox process models clustering by a random intensity field.

Misapplication

Assuming a Poisson model (complete spatial randomness) when data exhibit clustering or inhibition, or using intensity as if it were a probability density for individual realizations rather than an expected count measure.

Consequence

Correct use gives tools such as Campbell's formula, Palm distributions, likelihoods for inference, and tractable summary statistics (K-function, pair correlation) to quantify event interactions and to perform simulation and estimation.

Reversal

A continuous random field or deterministic point set: instead of discrete events, mass is spread continuously or points follow a fixed nonrandom pattern, removing randomness of counts in bounded sets.

Boundary

Covers models representing discrete event locations; excludes pure random measures with continuous mass density only, deterministic lattices, and phenomena better modeled by marked continuous fields unless marks and discreteness are explicit.

Semantic Tension

Tension with random measure theory: point processes are integer-valued random measures concentrated on countable sets, while the broader random measure concept permits non-integer mass and continuous components, leading to subtle distinctions in formalism and inference.

Synthesis

A Point Process is the mathematical object for modeling discrete events in space or time, organized by intensity and dependence structure; choices between Poisson, Cox, Gibbs or renewal classes capture independence, clustering, or inhibition behaviors in observed event patterns.