OT: Neural Networks and fp

[Walter Vaughan]

Howie wrote:

> Just curious if anyone has tried to implement a neural network using
> filePro.
> I would love to see even (especially) the simplest example.

I posed this question about two weeks ago, and JE took the thread down a 
  path from which it never returned.

There is a ton of information on n-networks on the internet, however 
it's next to impossible to get it distilled down to a level that I 
completely understand.

I am looking at what "nn" types call 'market basket' networks. There are 
several companies that specialize in this type of data mining, however 
they seem to be either really big, or really small.

I learned a lot reading .pdf's on basically neural networks for dummies.

Of course there are the links to open source sites...
http://www.jooneworld.com/

And demoware sites...
http://www.easynn.com/

And bigvendor sites...
http://www-106.ibm.com/developerworks/library/l-neural/

And hobby sites...
http://naramcheez.netfirms.com/tutorials.htm

However, from what I have read, Ken or whoever wrote the code to read 
the patterns from the fingerprint readers *had* to implement some type 
of neural network. Now if that code can be leveraged in the future as a 
function in filePro... well only time will tell.

I have noticed that a lot of college coursework revolves around JOONE, 
and since it can read JDBC exposed databases, excel spreadsheets, and 
html, I can imagine only more newly minted MBA's saying I can only make 
money for you with modern technology, and we've got to kill all 
databases that cannot be stored in the SAN.

On the other hand filepro already acts like a neural analyzer. *report 
makes two passes if asked, and stores information about the 1st pass as 
it does it's thing.

The framework is there.

--
Walter