From: Marcelo L. <mar...@gm...> - 2012-02-27 03:11:29
|
Hi All, I have a SOM trained by the Databionics ESOM tool. I also have a set of data that I need to run against this network to see which of the neurons of the network are activated by this set, but naturally many of the activated neurons are not the best matches. How can I determine the closest best match of the an activated neuron? I'm trying to create classes the way the help site says, but I'm not having success with that. Another option is to force the projection algorithm to only activate the best match neurons, I'm using the 'esomprj' tool, is there any way to do this? Thank you in advance. -- Marcelo Lotif. |