QuadraticDiscriminantAnalysis.DiscriminantFunctions Property

public ReadOnlyCollection<QuadraticDiscriminantFunction> DiscriminantFunctions { get; }

Public ReadOnly Property DiscriminantFunctions As ReadOnlyCollection(Of QuadraticDiscriminantFunction)
	Get

public:
property ReadOnlyCollection<QuadraticDiscriminantFunction^>^ DiscriminantFunctions {
	ReadOnlyCollection<QuadraticDiscriminantFunction^>^ get ();
}

member DiscriminantFunctions : ReadOnlyCollection<QuadraticDiscriminantFunction> with get

Property Value

The discriminant functions are the core of quadratic discriminant analysis. Each function corresponds to one class and computes the log posterior probability for a given observation belonging to that class. The function incorporates the class-specific mean vector, covariance matrix, and prior probability.

These functions can be used to classify new observations by evaluating each function and selecting the class with the highest score. The score includes both the Mahalanobis distance from the class centroid and terms accounting for the different covariance matrices between classes.

Reference