QuadraticDiscriminantAnalysis.GroupCovariances Property

public Matrix<double>[] GroupCovariances { get; }

Public ReadOnly Property GroupCovariances As Matrix(Of Double)()
	Get

public:
property array<Matrix<double>^>^ GroupCovariances {
	array<Matrix<double>^>^ get ();
}

member GroupCovariances : Matrix<float>[] with get

Property Value

The group covariances are the class-specific covariance matrices estimated from the training data. Each matrix captures the variance and correlation structure of the features within its corresponding class.

Unlike linear discriminant analysis which assumes equal covariance matrices across classes, QDA estimates separate covariance matrices for each class. This allows the model to capture class-specific feature relationships and variances, making it more flexible but requiring more parameters to be estimated.

This property is only available when the model is fitted. It throws an InvalidOperationException if the model is empty or deployed.

Reference