support for large sparse covariance matrices

[bruns01]

Some routines in the package make use of np.random.multivariate_normal(...) which samples from a multivariate normal distribution based on information given by the covariance matrix in the arguments to the call. In many cases of dynamic measurements long time series of combinations of time series lead to very large but (very) sparse covariance matrices. These cannot be handled efficiently by np.random.multivariate_normal(). A dedicated function fit to deal efficiently with scipy.sparse-matrices would be much appreciated!

[eichstaedtPTB]

This is very reasonable request, thanks! Actually, there are a couple of methods to address this issue and I am aware of a recently published paper from NPL colleagues on large scale covariance matrix compression methods. Therefore, I'll assign this feature request to the NPL contributors for PyDynamic to implement this.