home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!darwin.sura.net!wupost!sdd.hp.com!swrinde!zaphod.mps.ohio-state.edu!menudo.uh.edu!lobster!nuchat!texhrc!texhrc!ak45ldp
- From: ak45ldp@Texaco.com (Larry D. Pyeatt)
- Newsgroups: comp.ai.neural-nets
- Subject: Re: Back Propogation - ??????
- Keywords: Learning - Strange
- Message-ID: <1992Jul29.135042.3391@texhrc.uucp>
- Date: 29 Jul 92 13:50:42 GMT
- References: <1992Jul28.170531.3687@morgan.ucs.mun.ca>
- Sender: news@texhrc.uucp
- Organization: Texaco
- Lines: 20
- Nntp-Posting-Host: aisun
-
- In article <1992Jul28.170531.3687@morgan.ucs.mun.ca>, raghu@morgan.ucs.mun.ca (Raghu B) writes:
- ....
- |> I always find a bowl shaped nature in the neural behaviour.
- |> The error is at an acceptable level. Is bowl shape a characteristic
- |> of bp algorithm ? I am curious to know more about it.
- |>
-
- The NN should have been able to fit your data much better than
- your graph indicates. It is possible that you have too many
- neurons in your hidden layer, or that you do not have enough training
- data, or that you need to train for a longer period of time.
-
- You may notice some small nonlinearities in the NN response curve,
- but not on the order that your graph suggests.
-
- --
- Larry D. Pyeatt The views expressed here are not
- Internet : pyeatt@texaco.com those of my employer or of anyone
- Voice : (713) 975-4056 that I know of with the possible
- exception of myself.
-