NormalOptions Properties

Gets or sets a value indicating whether the covariance matrix to be estimated should be assumed to be diagonal.

Gets or sets a post processing step can be called after all component distributions have been fitted (or their .Fit() method has been called).

Gets or sets the regularization step to avoid singular or non-positive definite covariance matrices. Default is 0. Setting this property to a small constant like 1e-6 is more efficient than setting Robust to true.

Gets or sets whether the estimation function should allow non-positive definite covariance matrices by using the Singular Value Decomposition Function. Enabling this property can significantly increase learning times.

Gets or sets whether the normal distributions should have only a single, shared covariance matrix among all components in a mixture. Setting this property only has effect if the distributions are part of a MixtureT or MultivariateMixtureT