Difference between revisions of "Tips"

From SUMOwiki
Jump to navigationJump to search
Line 1: Line 1:
* The matlab neural network implementation is very slow! Try to avoid using crossvalidation but use test samples instead
+
* The matlab neural network implementation is very slow! Try to avoid using crossvalidation but use test samples instead (see [[FAQ|the FAQ entry]]).
* if you use the RBF neural network model type and you get a crash in "newrb" this is an error in the matlab implementation and not anything we can do about
+
* if you use the RBF neural network model type and you get a crash in "newrb" this is an error in the matlab toolbox implementation and not anything we can do about (a workaround is available on [[Contact|request]]).  This should be fixed by Matlab 7.5
 
* you can switch off adaptive sample selection if you do not specify a <SampleSelector> tag.  In this case all the available data will be used and only adaptive modeling will be done.  This only works with datasets.
 
* you can switch off adaptive sample selection if you do not specify a <SampleSelector> tag.  In this case all the available data will be used and only adaptive modeling will be done.  This only works with datasets.
 
* If you want to benchmark your computer for Matlab speed simply run "bench" in matlab
 
* If you want to benchmark your computer for Matlab speed simply run "bench" in matlab

Revision as of 16:39, 16 May 2007

  • The matlab neural network implementation is very slow! Try to avoid using crossvalidation but use test samples instead (see the FAQ entry).
  • if you use the RBF neural network model type and you get a crash in "newrb" this is an error in the matlab toolbox implementation and not anything we can do about (a workaround is available on request). This should be fixed by Matlab 7.5
  • you can switch off adaptive sample selection if you do not specify a <SampleSelector> tag. In this case all the available data will be used and only adaptive modeling will be done. This only works with datasets.
  • If you want to benchmark your computer for Matlab speed simply run "bench" in matlab