This PR adds an implementation of the minimum eigenvalue method, from doi:10.1080/16000870.2019.1696646, for conditioning of the covariance matrix. This will remove the use of the random walk model for the covariance matrix. Therefore, the default behaviour is now changed such that this model is not used by default.
It is still necessary to add a docstring for the cond_max input parameter that defines the covariance matrix's condition number after reconditioning and mention this in the documentation.
This PR adds an implementation of the minimum eigenvalue method, from doi:10.1080/16000870.2019.1696646, for conditioning of the covariance matrix. This will remove the use of the random walk model for the covariance matrix. Therefore, the default behaviour is now changed such that this model is not used by default.
It is still necessary to add a docstring for the
cond_maxinput parameter that defines the covariance matrix's condition number after reconditioning and mention this in the documentation.