Provide functions or vignette implementing structured correlation matrices

[grizant]

Implement these structured correlation matrices:

• Allow users to specify a correlation matrix consistent with ARMA (auto-regressive moving average) time series. See R package ts.extend (ARMA.var) for inspiration.
• Spiked correlation matrix

From Prof Ed Bedrick's commentary: "I think that your package would get more visibility if you provided some flexibility on structure.
As is you focus primarily on settings where the correlation can be arbitrary but many problems
are such that that correlation matrix has more structure - so giving people some options on
specifying certain high-dimensional correlation structures automatically (or just
illustrate how easy to do in R!) would be great.

Two simple enhancements are that you could
(a) allow users to specify that they want a "correlation" structure (Pearson, Spearman, Kendall)
consistent with an ARMA (auto-regressive moving average) time series structure - in the
R package ts.extend (using ARMA.var) you can generate the Pearson correlation
structure for an ARMA model - if one views the specification as a Spearman structure then
you can generate a time series with arbitrary marginals that has that Spearman structure!
(b) lots of high-dimensional multivariate methods assumes a "spiked covariance matrix" - it
seems that you can easily program certain versions and allow people to pick these
structures for either Kendall, Pearson and Spearmen and take it from there

Simple changes like this - either illustrate their use in examples or as features of the package
would be a good idea"