Loading ...
Sorry, an error occurred while loading the content.

1785EasyNN secrets?

Expand Messages
  • P. Dickens
    Jun 20, 2013
    • 0 Attachment
      Hi Marjola,

      On reading your enquiry I imediately thought of John
      Mcenroe............ " You can't be serious?"
      Just look at the number of editions of ENN! How on God's earth do you
      think it would have got that far if, " It worked too well to be trusted?"

      P. Dickens


      From:
      "marjola24" <marjola24@...>


      * Save Addresses
      <http://email10.orange.co.uk/webmail/en_GB/read.html?FOLDER=UF_NN%2FENN&IDMSG=97&check=&ORIGIN=PERSONNAL_FOLDER&SORTBY=1#>

      To:
      easynn@yahoogroups.com
      Date: Jun 19 2013, 04:39 PM
      Subject:
      EasyNN secrets?


      Hi Guys, I'm evaluating EasyNN (of a friend of mine) and it worksO so
      well that I'm thinking to purchase it.
      The problem is that it works too well!
      I compared EasyNN with a software written by myself and It finds the
      results just about after few cycles, whereas my software never reaches
      to a solution (3 inputs, 10 outputs for 27 samples)

      The problem is that the algorithms should be the same and also the
      procedure for data normalization:

      What's the secret of EasyNN? It works too well to be trusted?

      Thanks for helping me,
      Marjola.


      [Non-text portions of this message have been removed]
    • Show all 2 messages in this topic