home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!sun-barr!ames!network.ucsd.edu!sdcc12!cs!demers
- From: demers@cs.ucsd.edu (David DeMers)
- Newsgroups: comp.ai.neural-nets
- Subject: Re: function approximation with neural nets : what is right ?
- Message-ID: <37892@sdcc12.ucsd.edu>
- Date: 10 Sep 92 18:53:09 GMT
- References: <1992Sep10.153652.12503@noose.ecn.purdue.edu>
- Sender: news@sdcc12.ucsd.edu
- Organization: =CSE Dept., U.C. San Diego
- Lines: 27
- Nntp-Posting-Host: beowulf.ucsd.edu
-
- In article <1992Sep10.153652.12503@noose.ecn.purdue.edu> kavuri@lips2.ecn.purdue.edu (Surya N Kavuri ) writes:
-
-
- > In using a neural net to approximate a function, it is done with
- > out any a priori information on what class of functions we are
- > looking for. For example, if I know the function is some
- > polynomial, there is no way to impose this condition on the net.
-
- > Is not this a disadvantage ?
-
- It's the "disadvantage" accruing to all non-parametric methods.
- It's more of using the wrong tool. If you know that the
- true function is a polynomial of a certain order, you will do
- better by just estimating the parameters. But if you *think* it's
- a polynomial, and you are wrong (about the order, at least),
- then even the best approximation you make will be poor.
-
- So use a parametric technique if you have a good idea of the
- model class, and use a non-parametric technique if you have
- no a priori information.
-
-
- --
- Dave DeMers ddemers@UCSD demers@cs.ucsd.edu
- Computer Science & Engineering C-014 demers%cs@ucsd.bitnet
- UC San Diego ...!ucsd!cs!demers
- La Jolla, CA 92093-0114 (619) 534-0688, or -8187, FAX: (619) 534-7029
-