Generating random \(\mathrm{AR}(p)\) and \(\mathrm{MA}(q)\) Toeplitz correlation matrices (Q968505): Difference between revisions

The authors propose methods for generating random \((p+1)\times (p+1)\) Toeplitz correlation matrices that are consistent with a causal \(\text{AR}(p)\) Gaussian time series model. Also, methods are proposed for generating \((q+1)\times (q+1)\) Toeplitz correlation matrices that are consistent with an invertible \(\text{MA}(q)\) Gaussian time series model. The random generating methods are useful for models with a structured Toeplitz matrix as a parameter.