Multivariate Class

Contains methods for analyzing multivariate samples.

Inheritance Hierarchy

System.Object
Meta.Numerics.Statistics.Multivariate

Namespace: Meta.Numerics.Statistics
Assembly: Meta.Numerics (in Meta.Numerics.dll) Version: 4.1.4

Syntax

C++

Copy

public static class Multivariate

<ExtensionAttribute>
Public NotInheritable Class Multivariate

[ExtensionAttribute]
public ref class Multivariate abstract sealed

[<AbstractClassAttribute>]
[<SealedAttribute>]
[<ExtensionAttribute>]
type Multivariate =  class end

The Multivariate type exposes the following members.

Methods

	Name	Description
	MeansClustering	Compute k-means clusters of the data.
	MultiLinearLogisticRegression(IReadOnlyList<Boolean>, IReadOnlyDictionary<String, IReadOnlyList<Double>>)	Performs a multivariate linear logistic regression on the named columns.
	MultiLinearLogisticRegression(IReadOnlyList<Boolean>, IReadOnlyList<IReadOnlyList<Double>>)	Performs a multivariate linear logistic regression on the listed columns.
	MultiLinearLogisticRegression(IReadOnlyList<Boolean>,IReadOnlyList<Double>[])	Performs a multivariate linear logistic regression.
	MultiLinearRegression(IReadOnlyList<Double>, IReadOnlyDictionary<String, IReadOnlyList<Double>>)	Performs a multivariate linear regression on the named columns.
	MultiLinearRegression(IReadOnlyList<Double>, IReadOnlyList<IReadOnlyList<Double>>)	Performs a multivariate linear regression on the listed columns.
	MultiLinearRegression(IReadOnlyList<Double>,IReadOnlyList<Double>[])	Performs a multivariate linear regression.
	PrincipalComponentAnalysis(IReadOnlyList<IReadOnlyList<Double>>)	Performs a principal component analysis of the multivariate sample.
	PrincipalComponentAnalysis(IReadOnlyList<Double>[])	Performs a principal component analysis of the columns.
	RawMoment(IReadOnlyList<IReadOnlyList<Double>>, IReadOnlyList<Int32>)	Computes the multivariate raw moment specified by the listed powers.
	RawMoment(IReadOnlyList<IReadOnlyList<Double>>,Int32[])	Computes the specified multivariate raw moment.

Top

Remarks

A multivariate sample is one in which each element of the sample consists of measurements of multiple quantities. For example, a survey in which we record the sex, income, education level, and race of each participant is a multivariate sample. Note that the types of each quantity measured need not be the same.

One common form of statistical analysis of a multivariate sample is to try to predict one variable on the basis of some of the others. To do this you can use MultiLinearRegression(IReadOnlyList<Double> ,IReadOnlyList<Double> []) or, if the dependent variable is Boolean, MultiLinearLogisticRegression(IReadOnlyList<Boolean>,IReadOnlyList<Double>[]).

Another common operation of multivariate data is to try to organize the observations into clusters, which you can do using MeansClustering(IReadOnlyList<IReadOnlyList<Double>>, Int32).

Instead of trying to reduce all the observations to a few important representatives, you can try to reduce all the variables to a few important directions using PrincipalComponentAnalysis(IReadOnlyList<IReadOnlyList<Double>>).

Reference

Meta.Numerics.Statistics Namespace