LIM (version 1.4.7.1)

Xsample: Generates a random sample of the unknowns for a linear inverse problem

Description

Given an inverse input list, randomly samples the unknowns, using an MCMC method

Usage

Xsample(lim, exact = NULL, ...)

Value

a 2-columned vector containing the minimum (column 1) and maximum (column 2) of each unknown.

Arguments

lim

a list that contains the linear inverse model specification, as generated by function setup.limfile.

exact

if not NULL, a vector containing the numbers of the equations to be solved exactly; if NULL, all equations are considered exact.

...

extra parameters passed to function xsample from packagelimSolve.

Author

Karline Soetaert <[e-mail]>

Details

For overdetermined LIM problems, the inverse equations may be split up in equations which have to be exactly met and other equations which have to be approximated.

exact is a vector with the exact equations

The default settings of xsample will often not do. For instance, the default consists of 3000 iterations (iter) and a jump length of jmp of 0.1. You may need to increase one of those to ensure that the entire solution space has been adequately sampled.

References

Van den Meersche K, Soetaert K, Van Oevelen D (2009). xsample(): An R Function for Sampling Linear Inverse Problems. Journal of Statistical Software, Code Snippets, 30(1), 1-15.

http://www.jstatsoft.org/v30/c01/

See Also

Varranges which estimates the ranges of inverse variables

Plotranges to plot the ranges

function xsample from packagelimSolve

Examples

# sample solution space
  xs <- Xsample(LIMRigaAutumn, iter = 500, jmp = 5)
  # remove flows that are invariable (sd=0)
  xs <- xs[ ,-which(apply(xs, 2, sd) == 0 )]
  #pairs plot
  pairs(xs, gap = 0, pch = ".", upper.panel = NULL)