Neural Net Software Recom.

Discussion in 'Trading Software' started by cosmic, Apr 7, 2003.

  1. cosmic

    cosmic

    Hi,

    I am searching a very good neural net software with an intraday interface to TradeStation 2000i.

    So far I have come up with the following products:

    1. BrainCel from JRC

    2. Ward Systems NeuralShell Day Trader Prof. (although lacking intraday interface to TS at the moment)

    3. Safir X from SirTrade.

    What else could be a possibility? I should have plenty of room for customizing needs. A definitve must have is intraday relearning with user defined intervall setting.

    I am using BrainCel right now, but for future (bigger) projects it is not advanced enough. Just my opinion, for the basics it is very good & learns amazingly fast.

    Thank you for any hints.... :)

    Cosmic
     
  2. "BrainCel" -LMAO I like that :D
     
  3. CalTrader

    CalTrader Guest

    We looked at all of these and many other related ideas that have commercial implementations. At the end of the day our own technologies were the only valid solution and we bit the bullet and did the custom work. A bit of advice: be very careful with the existing packages and solutions that are available - including others that are not merely TS interface products. In our testing we found issues of various degrees of severity with the existing commercial products. If you needs are very modest these may have some utility .....
     
  4. BzTrader

    BzTrader

    CallTrader

    Which is your "own technologies" ?

    Thanks

    BzTrader:)
     
  5. cosmic

    cosmic

    for all your comments :D

    Seems like very few people are actually working with NNs in realtime-reallife? If this is not the case, please tell what other good products are out there! :)

    Thank you,
    COSMIC :)
     
  6. CalTrader

    CalTrader Guest

    Stuff we developed over several years on things related to neural nets - supersets of these ideas. ....

    The primary issue with neural nets and related ideas is that very little progress has been made in understanding these areas over the last couple of decades: there are a lot of papers and some progress but by and large most of the useful information is not published. Once you actually start building large scale structures you begin to understand and interpret all the hype surrounding the field - even by the prominant researchers in the field. Techniques that are actually useful require discovery and dont appear to be published .....