The Setting
Let be a natural number, let be a real number with , and let be independent standard Gaussian variables. The stationary Gaussian autoregressive chain with coefficient starts at and moves by (1)
Given coefficients for the coordinates of an index set , the vector chain with coefficients is the family of states , , whose coordinate is the stationary Gaussian autoregressive chain with coefficient , all driven by a noise array with and , distributed as the standard Gaussian on . Time is discrete: in the estimates on uniform Gaussian survival, this chain, observed at the steps , takes the place of the stationary Ornstein-Uhlenbeck process.
The Autoregressive Row
The autoregressive row of a real coefficient is the array of reals , indexed by natural numbers (the step) and (the noise index), defined by recursion on the step: (2)✦ where is when holds and otherwise. For , the entry is the weight of the noise in the state of (1), so that for every .
The Claim
Definition 1. Let be real coefficients, a natural number and a step with . The autoregressive evaluation at step is the real matrix with rows indexed by and columns indexed by the pairs , whose entries are (3)✦ Applied to a noise array , it gives the state at step of the vector chain with coefficients .
Theorem 1. Let be a real number with , and let be natural numbers. Then the states and of the chain with coefficient satisfy (4)✦
In the paper, the result reads: the stationary Ornstein-Uhlenbeck process solves , where is a positive definite matrix, with standard Gaussian and a Brownian motion, and .
What Is Checked
The definition (3) is stated over the autoregressive row (2) of each coefficient , and the covariance identity (4) over the row of a single coefficient with , at steps .