SVC_Q

Full name Microsoft.VisualBasic.MachineLearning.SVM.SVC_Q Assembly Microsoft.VisualBasic.MachineLearning Members 4

Q matrices for various formulations

01 Syntax

Microsoft.VisualBasic.MachineLearning.SVM.SVC_Q

02 Methods

Name	Overloads	Summary
.ctor	1	Create the C-SVC / nu-SVC formulation of the kernel matrix.
GetQ	1	Request a column of the kernel matrix, the value is multiplied by the sign of the class labels: Q(i, j) = y_i * y_j * K(x_i, x_j).
GetQD	1	Gets the diagonal elements of the kernel matrix.
SwapIndex	1	Swap the position of the two samples in the kernel matrix, the cached columns, the class labels and the diagonal elements are swapped at the same time.

03 Members

method .ctor #

#ctor(Problem, Parameter, SByte())

Create the C-SVC / nu-SVC formulation of the kernel matrix.

Parameters

Name	Type	Description
`prob`	Problem	The training data.
`param`	Parameter	The training parameters.
`y_`	`SByte()`	The signed class label (+1/-1) of each training sample.

method GetQ #

GetQ(Int32, Int32)

Request a column of the kernel matrix, the value is multiplied by the sign of the class labels: Q(i, j) = y_i y_j K(x_i, x_j).

Parameters

Name	Type	Description
`i`	`Int32`	The zero based index of the target row.
`len`	`Int32`	The number of the elements that will be requested.

Returns

An array which contains the first len elements of the target column.

method GetQD #

GetQD

Gets the diagonal elements of the kernel matrix.

Returns

An array which contains the diagonal elements Q(i, i).

method SwapIndex #

SwapIndex(Int32, Int32)

Swap the position of the two samples in the kernel matrix, the cached columns, the class labels and the diagonal elements are swapped at the same time.

Parameters

Name	Type	Description
`i`	`Int32`	The zero based index of the first sample.
`j`	`Int32`	The zero based index of the second sample.