home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai.neural-nets
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!decwrl!sdd.hp.com!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!wupost!gumby!yale!yale.edu!jvnc.net!nuscc!nusunix3.nus.sg!isc30221
- From: isc30221@nusunix1.nus.sg (CHONG SOON FAI)
- Subject: Kohonen (self-organizing) maps parameters
- Message-ID: <1992Nov6.012107.27936@nuscc.nus.sg>
- Sender: usenet@nuscc.nus.sg
- Organization: National University of Singapore
- X-Newsreader: TIN [version 1.1 PL6]
- Date: Fri, 6 Nov 1992 01:21:07 GMT
- Lines: 14
-
- hi! I m doing a project on Kohonen (self-organizing)
- semantic/topological maps. I am simulating the neural net in software
- in software.
-
- Does anyone have any suggestions as to what neural net parameters (eg
- size of network, number of training vectors, size of neighbourhood, what
- function to use for the learning coefficient etc) might be useful in
- getting the net to converge?
-
- Thanks!
-
-
- SF
- "my first post to internet"
-