Warp: The 'Warp' class object

In mu2013/KGode: Kernel Based Gradient Matching for Parameter Inference in Ordinary Differential Equations

Description

This class provide the warping method which can be used to warp the original signal to sinusoidal like signal.

Format

R6Class object.

Value

an R6Class object which can be used for doing interpolation using reproducing kernel Hilbert space.

Methods

warpsin(len ,lop,p0,eps)

This method is used to warp the initial interpolation into a sinusoidal shape.

slowWarp(lens,peod,eps)

This method is used to find the optimised initial hyper parameters for the sigmoid basis function for each ode states.

warpLossLen(par,lam,p0,eps)

This method is used to implement the loss function for warping. It is called by the 'warpSin' function.

Public fields

y

matrix(of size n_s*n_o) containing observation.

t

vector(of length n_o) containing time points for observation.

b

vector(of length n_o) containing coefficients of kernel or basis functions.

lambda

scalar containing the weighting parameter for penalising the length of warped time span.

ker

kernel class object containing sigmoid basis function.

Methods

Public methods

• Warp$new()

• Warp$greet()

• Warp$showker()

• Warp$warpLoss()

• Warp$warpLossLen()

• Warp$warpSin()

• Warp$slowWarp()

• Warp$clone()

---

Method new()

Usage

Warp$new(y = NULL, t = NULL, b = NULL, lambda = NULL, ker = NULL)

---

Method greet()

Usage

Warp$greet()

---

Method showker()

Usage

Warp$showker()

---

Method warpLoss()

Usage

Warp$warpLoss(par, len, p0, eps)

---

Method warpLossLen()

Usage

Warp$warpLossLen(par, lam, p0, eps)

---

Method warpSin()

Usage

Warp$warpSin(len, lop, p0, eps)

---

Method slowWarp()

Usage

Warp$slowWarp(lens, p0, eps)

---

Method clone()

The objects of this class are cloneable with this method.

Usage

Warp$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Author(s)

Mu Niu, [e-mail]

mu2013/KGode documentation built on June 23, 2020, 8:04 p.m.