Theory
An AR(1) time series satisfies
Xn + 1 =
ρ Xn
+ Yn
where the Yn
are IID
Normal(0, τ2).
The lag k autocovariance is
γk
=
ρk τ2 ⁄ (1 - ρ2)
The asymptotic variance in the central limit theorem is
σ2
=
γ0 (1 + ρ) ⁄ (1 - ρ)
The approximate, large k, variance of the empirical
estimates of the γk is approximately
1 ⁄ n times
γ02
+
2 ∑j = 1∞
γj2
Simulation
Batch Means
Overlapping Batch Means
Behavior of Empirical Autocovariances
Estimators Using Reversibility
Γk
=
γ2 k
+
γ2 k + 1
is strictly positive, strictly decreasing, strictly convex function of
k.
Truncate sequence at first negative estimate of Γk.
Gives initial positive sequence estimator.
Take maximimal decreasing function that fits under above.
Gives initial decreasing sequence estimator.
Take maximimal decreasing convex function that fits under above.
Gives initial convex sequence estimator.
The Old Way with Lots of Code
The New Way with New Function