[EMAIL PROTECTED] (akhan) wrote 
"Is there any statistical metod which can be applied to test whether a
non-linear model fit a dataset well or significantly?"

One can assess how far off the model is expected to be from the
underlying process through any of a number of resampling procedures,
regardless of the modeling process (or even if you are reading Tarot
cards!).

The simplest such method is holdout testing: fit your model using some
randomly selected fraction of the available data and test it (using
whatever criteria is of interest, such as mean absolute error) on the
remaining, distinct fraction of the data.  Pro: simple, easy to do on
a computer.  Con: needing to know how much data to holdout.

With small data sets, simple holdout testing can be unstable, and one
would consider procedures such as k-fold cross-validation (including
leave-one-out), bootstrapping, etc.  Search on any of these terms
online, or examine the relevant material in Weiss and Kulikowski's
"Computer Systems That Learn".

-Will Dwinnell
http://will.dwinnell.com
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to