MachineLearningknn Method

public static RetArray<long> knn(
	InArray<double> Samples,
	InArray<double> Neighbors,
	int k = 10,
	DistanceMetrics metric = DistanceMetrics.Euclidian_L2,
	double minkowski_parameter = 2,
	bool unstable_error = true
)

Public Shared Function knn ( 
	Samples As InArray(Of Double),
	Neighbors As InArray(Of Double),
	Optional k As Integer = 10,
	Optional metric As DistanceMetrics = DistanceMetrics.Euclidian_L2,
	Optional minkowski_parameter As Double = 2,
	Optional unstable_error As Boolean = true
) As RetArray(Of Long)

Dim Samples As InArray(Of Double)
Dim Neighbors As InArray(Of Double)
Dim k As Integer
Dim metric As DistanceMetrics
Dim minkowski_parameter As Double
Dim unstable_error As Boolean
Dim returnValue As RetArray(Of Long)

returnValue = MachineLearning.knn(Samples, Neighbors, 
	k, metric, minkowski_parameter, unstable_error)

Parameters

Samples

Type: ILNumericsInArrayDouble
Samples matrix, samples in columns, the number of rows (dimensionality) must match the number of rows in Neighbors.

Neighbors

Type: ILNumericsInArrayDouble
Matrix of training samples/ neighbors, this will be searched for matching points, rows: dimensionality, columns: number of points.

k (Optional)

Type: SystemInt32
[Optional] Number of neighbors to return, k must lay in range: 0 <= k < neighbors.D[1]; default: 1.

metric (Optional)

Type: ILNumerics.ToolboxesDistanceMetrics
[Optional] Distance metric, one out of the [!:ILNumerics.Toolboxes.MachineLearning.DistanceMetrics] enumeration. Supported are: Euclidian_L2,Manhattan_L1, Minkowski, Cosine, Pearsons and Hamming distances; default: 'Euclidian_L2'.

minkowski_parameter (Optional)

Type: SystemDouble
[Optional] Exponent for minkowski distance; default: 2.

unstable_error (Optional)

Type: SystemBoolean
[Optional] For cosine and pearson distances: if some samples lead to numerical instabilities, an exception is generated; default: true.

Return Value

[ILNumerics Machine Learning Toolbox]

Reference