make.findif {CollocInfer}

Finite Difference Functions

Description

Returns a list of functions that calculate finite difference derivatives.

Usage

make.findif.ode()

make.findif.loglik()

make.findif.var()

Details

All these functions require the sepcification of more$eps to give the size of the finite differencing step. They also require more to specify the original object (ODE right hand side functions, definitions of lik and proc objects).

Value

A list of functions that calculate the derivatives via finite differencing schemes.

make.findif.ode

calculates finite differences of a transform.

make.findif.loglik

returns the finite differences to a calculated log likelihood; used within lik objects, or as more arguments to Cproc or Dproc.

make.findif.var

finite difference approximations to variances; mostly used in the Multinorm functions.

See Also

LS.setup, multinorm.setup

Examples

# Sum of squared errors with finite differencing to get right-hand-side derivatives

proc = make.SSEproc()
proc$more = make.findif.ode()


# Finite differencing for the log likelihood

lik = make.findif.loglik()
lik$more = make.SSElik()


# Multivariate normal transitions with finite differencing for mean and variance functions

lik = make.multinorm()
lik$more = c(make.findif.ode,make.findif.var)

# Finite differencing for transition density of a discrete time system

proc = make.Dproc()
proc$more = make.findif.loglik()