Link Search Menu Expand Document
matrix-ts

Multivariate overview

Useful matrix operations for multivariate samples.

Added in v1.0.0

Table of contents

Model

MultivariateSample (type alias)

An observation is a (free) collection of data vectors, where N is the number of variables associated with the observation

Signature

export type MultivariateSample<N> = RNEA.ReadonlyNonEmptyArray<V.Vec<N, number>>

Example

// Given three subjects whose measurements are height, weight, shoe size, we can construct a correlation matrix as follows:

import * as V from '@jacob-alford/matrix-ts/Vector'
import * as MStat from '@jacob-alford/matrix-ts/Multivariate'

const obs1 = V.fromTuple([167, 63, 8])
const obs2 = V.fromTuple([200, 94, 12])
const obs3 = V.fromTuple([160, 65, 9])
const observations: MStat.MultivariateSample<3> = [obs1, obs2, obs3]

// This determines how mutually correlated the different variables are
MStat.correlation(observations)

Added in v1.0.0

Multivariate Statistics

correlation

The correlation matrix where each index: i, j represent the correlation between those two random variables

Signature

export declare const correlation: <N extends number>(
  s: RNEA.ReadonlyNonEmptyArray<V.Vec<N, number>>
) => M.Mat<N, N, number>

Added in v1.0.0

covariance

The covariance matrix where each index: i, j represent the variance between each random variable of the Multivariatesample

Signature

export declare const covariance: <N extends number>(
  s: RNEA.ReadonlyNonEmptyArray<V.Vec<N, number>>
) => M.Mat<N, N, number>

Added in v1.0.0

deviation

The difference in observation from a mean of MultivariateSample

Signature

export declare const deviation: <N extends number>(
  s: RNEA.ReadonlyNonEmptyArray<V.Vec<N, number>>
) => RNEA.ReadonlyNonEmptyArray<V.Vec<N, number>>

Added in v1.0.0

mean

The average value over different variables

Signature

export declare const mean: <N extends number>(s: RNEA.ReadonlyNonEmptyArray<V.Vec<N, number>>) => V.Vec<N, number>

Added in v1.0.0