Periodic

Defined in: ds/src/ml/kernels/periodic.js:15

Abstract base class for GP kernels

Extends

• Kernel

Constructors

Constructor

new Periodic(lengthScale?, period?, variance?): Periodic

Defined in: ds/src/ml/kernels/periodic.js:21

Parameters

lengthScale?

number = 1.0

Length scale (default: 1.0)

period?

number = 1.0

Period length (default: 1.0)

variance?

number = 1.0

Signal variance (default: 1.0)

Returns

Periodic

Overrides

Kernel.constructor

Properties

lengthScale

lengthScale: number

Defined in: ds/src/ml/kernels/periodic.js:23

---

period

period: number

Defined in: ds/src/ml/kernels/periodic.js:24

---

variance

variance: number

Defined in: ds/src/ml/kernels/periodic.js:25

Methods

call()

call(X1, X2?): Matrix

Defined in: ds/src/ml/kernels/base.js:30

Compute covariance matrix between sets of points

Parameters

X1

any

First set of points (n1 x d)

X2?

any = null

Second set of points (n2 x d). If omitted, computes K(X1, X1)

Returns

Matrix

Covariance matrix (n1 x n2)

Inherited from

Kernel.call

---

clone()

clone(): Kernel

Defined in: ds/src/ml/kernels/base.js:76

Clone the kernel with the same parameters

Returns

Kernel

New kernel instance

Inherited from

Kernel.clone

---

compute()

compute(x1, x2): number

Defined in: ds/src/ml/kernels/periodic.js:28

Compute covariance between two points

Parameters

x1

any

First point

x2

any

Second point

Returns

number

Covariance value

Overrides

Kernel.compute

---

getParams()

getParams(): object

Defined in: ds/src/ml/kernels/periodic.js:41

Get kernel hyperparameters

Returns

object

Hyperparameters

lengthScale

lengthScale: number

period

period: number

variance

variance: number

Overrides

Kernel.getParams

---

setParams()

setParams(params): void

Defined in: ds/src/ml/kernels/periodic.js:49

Set kernel hyperparameters

Parameters

params

New parameters

lengthScale

any

period

any

variance

any

Returns

void

Overrides

Kernel.setParams