Contamination (statistics)

Statistical contamination replaces part of an ideal distribution or sample by arbitrary observations.

The Kolmogorov $\epsilon$-neighborhood of a cumulative distribution function $F$ consists of all distribution functions $G$ satisfying

The epsilon-contamination neighborhood of $P$ is the class $\{(1-ϵ)P+ϵQ:Q\text{ is a probability distribution}\}$.

Huber contamination is the epsilon-contamination model in which an arbitrary distribution contributes an unknown fraction of the observations.