Data PDF of stationary process

Let $ {\bf R}_N$ be the $ N\times N$ covariance matrix of a stationary process. Since the process is stationary, $ {\bf R}_N$ will be symmetric and Toeplitz. The PDF may be written as

$\displaystyle \log p({\bf x}) = -\frac{N}{2} \log (2\pi) - \frac{1}{2}\log \ver...
...det}({\bf R}_N)\vert - \frac{1}{2} {\bf x}^\prime \; {\bf R}_N^{-1} \; {\bf x},$ (9.6)

where $ {\bf R}_N$ is the $ N\times N$ covariance matrix of an AR process.



Baggenstoss 2017-05-19