bbolker at gmail.com
Thu Mar 22 23:22:40 CET 2012
Berend Hasselman <bhh <at> xs4all.nl> writes:
> On 22-03-2012, at 18:37, mhimanshu wrote:
> > How to find best parameter values for deSolve function??
> I assume you mean the deSolve package which has many options.
> I would suggest predicting which way the wind will blow the
> day after tomorrow.
I think this is a less-well-worded repeat of this question:
which is not unreasonable, but no-one's gotten around to answering it
(Terse answer: construct a function that takes the parameters of
the ODE, and the initial conditions if you don't know them but
estimate them at the same time, solves the ODE numerically, and
computes a goodness-of-fit score (actually a badness of fit score,
such as the sum of squares or negative log-likelihood), and use
it as the objective function in optim().)
More information about the R-help