Читать книгу Computational Statistics in Data Science - Группа авторов - Страница 131

Although both and may be used in MCMC, a third alternative arises due to the correlation in the Markov chain. A relative‐standard deviation sequential stopping rule terminates the simulation when the Monte Carlo variability (as measured by the volume of the confidence region) is small compared to the underlying variability inherent to the problem . That is,

If this rule is used for IID Monte Carlo, then in Equation (2) is , and for some other (deterministic) . For MCMC, this sequential stopping rule connects directly to the concept of effective sample size [26]. That is, stopping at is equivalent to stopping when

(7)

Thus, simulation is terminated when the number of effective samples is larger than the lower bound in Equation (7). Effective sample size measures the number of equivalent IID samples that would produce equivalent variability in . Terminating simulation using Equation (7) is intuitive and easy to implement in MCMC sampling once appropriate estimators of and have been obtained.