home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.graphics.visualization
- Path: sparky!uunet!darwin.sura.net!Sirius.dfn.de!fauern!fauna!mskuhn
- From: mskuhn@immd4.informatik.uni-erlangen.de (Markus Kuhn)
- Subject: Re: Neural Networks and HyperPlanes
- References: <1992Jul20.173438.12081@cis.uab.edu>
- Message-ID: <BrqABF.K9I@immd4.informatik.uni-erlangen.de>
- Organization: CSD., University of Erlangen, Germany
- Date: Tue, 21 Jul 1992 07:33:15 GMT
- Lines: 26
-
- vachha@cisa.cis.uab.edu (Rustom (Yuppy) Vachha) writes:
-
- >This implementation currently restricts visualization of
- >networks with only 2 input nodes. If we have three input nodes, the
- >hyperplanes will be PLANES and with 4 input nodes, ??? .
-
- >HENCE WE NEED TO DEFINE A STRATEGY TO CONVERT N-dimensional DATA INTO 2 D.
- > How do we represent N dim hyperplanes?
-
- Representing 3-d hyperplanes in a 4-d space is still quite easy: Make
- an 3-d animation with moving points and hyperplanes. Time is the fourth
- dimension, i.e. a hyperplane is
-
- ax + by + cz + dt + e = 0
-
- You get another few dimensions by using colors, but then the person
- looking at the images has to be quite experienced.
-
- Have fun ...
-
- Markus
-
- --
- Markus Kuhn, Computer Science student -=-=- University of Erlangen, Germany
- Internet: mskuhn@immd4.informatik.uni-erlangen.de | X.500 entry available
- ------------------ ISO? Nicht immer. Aber immer ÷fter! -------------------
-