fisher (Model)

Approximate Fisher information matrix in frequency domain

Syntax

[F, FF, Delta, Freq] = fisher(M, NPer, PList, ...)

Input Arguments

M [ model ]

Solved model object.

NPer [ numeric ]

Length of the hypothetical range for which the Fisher information will be computed.

PList [ cellstr ]

List of parameters with respect to which the likelihood function will be differentiated.

Output Arguments

F [ numeric ]

Approximation of the Fisher information matrix.

FF [ numeric ]

Contributions of individual frequencies to the total Fisher information matrix.

Delta [ numeric ]

Kronecker delta by which the contributions in Fi need to be multiplied to sum up to F.

Freq [ numeric ]

Vector of frequencies at which the Fisher information matrix is evaluated.

Options

CheckSteady [ true | false | cell ]

Check steady state in each iteration; works only in non-linear models.

Deviation [ true | false ]

Exclude the steady state effect at zero frequency.

Exclude [ char | cellstr | empty ]

List of measurement variables that will be excluded from the likelihood function.

Percent [ true | false ]

Report the overall Fisher matrix F as Hessian w.r.t. the log of variables; the interpretation for this is that the Fisher matrix describes the changes in the log-likelihood function in reponse to percent, not absolute, changes in parameters.

Progress [ true | false ]

Display progress bar in the command window.

Solve [ true | false | cellstr ]

Re-compute solution in each differentiation step; you can specify a cell array with options for the solve() function.

Steady [ true | false | cell ]

Re-compute steady state in each differentiation step; if the model is non-linear, you can pass in a cell array with opt used in the steady() function.

Description

Example