makeDerivs: Utility functions for creating Gradient- and Hessian-like...

Description Usage Arguments Details Value Author(s) References Examples

Description

These are three different utility functions that create matrices containing the symbolic partial derivatives of first (makeGrad) and second (makeHess) order and a function for evaluating these matrices in an environment. The evaluations of the symbolic derivatives are used within the propagate function to calculate the propagated uncertainty, but these functions may also be useful for other applications.

Usage

1
2
3
makeGrad(expr, order = NULL)
makeHess(expr, order = NULL)
evalDerivs(deriv, envir)

Arguments

expr

an expression, such as expression(x/y).

order

order of creating partial derivatives, i.e. c(2, 1). See 'Examples'.

deriv

a matrix returned from makeGrad or makeHess.

envir

an environment to evaluate in. By default the workspace.

Details

Given a function f(x_1, x_2, …, x_n), the following matrices containing symbolic derivatives of f are returned:

makeGrad:

\nabla(f) = ≤ft[\frac{\partial f}{\partial x_1}, …, \frac{\partial f}{\partial x_n}\right]

makeHess:

H(f) = ≤ft[ \begin{array}{cccc} \frac{\partial^2 f}{\partial x_1^2} & \frac{\partial^2 f}{\partial x_1\,\partial x_2} & \cdots & \frac{\partial^2 f}{\partial x_1\,\partial x_n} \\ \frac{\partial^2 f}{\partial x_2\,\partial x_1} & \frac{\partial^2 f}{\partial x_2^2} & \cdots & \frac{\partial^2 f}{\partial x_2\,\partial x_n} \\ \vdots & \vdots & \ddots & \vdots \\ \frac{\partial^2 f}{\partial x_n\,\partial x_1} & \frac{\partial^2 f}{\partial x_n\,\partial x_2} & \cdots & \frac{\partial^2 f}{\partial x_n^2} \end{array} \right]

Value

The symbolic or evaluated Gradient/Hessian matrices.

Author(s)

Andrej-Nikolai Spiess

References

The Matrix Cookbook (Version November 2012).
Petersen KB & Pedersen MS.
http://ais.informatik.uni-freiburg.de/teaching/ws12/mapping/pdf/imm3274.pdf

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
EXPR <- expression(a^b + sin(c))
ENVIR <- list(a = 2, b = 3, c = 4)

## First-order partial derivatives: Gradient.
GRAD <- makeGrad(EXPR) 

## This will evaluate the Gradient.
evalDerivs(GRAD, ENVIR)

## Second-order partial derivatives: Hessian.
HESS <- makeHess(EXPR) 

## This will evaluate the Hessian.
evalDerivs(HESS, ENVIR)

## Change derivatives order.
GRAD <- makeGrad(EXPR, order = c(2,1,3)) 
evalDerivs(GRAD, ENVIR)

anspiess/propagate documentation built on May 14, 2019, 3:09 a.m.