Hi Tom
I'm not trying to ignore you. I have your previous message and was
going to email you a copy of the code when it was 'done'. Basically all
this is in response to a message by Bill Kerr as to whether one could model
a perceptron in Logo and possibly tie it to Lego-Logo. A perceptron is
just a neural net.
I replied to Bill that it was probably possible in Logo and have
implemented a few things. One problem is that some of the Logos I work
with are slow for the needed training of the net. A first cut then was
to do the training 'off-line' and let the Logo net do the classification.
However, just for fun, I now, for a certain situation, have some training
code in Logo. However, it is nice to use arrays and it would be nice to
be able to tell the color of a pixel. BUT I don't want to make the code
inaccesssible to other Logo folks and wanted to know what else was out
there.
My concern is that the code is generally usable. When I am more or less
satisfied that I have it near that level, I'll be sure to email you a copy.
Ed Wall
---------------------------------------------------------------
Please post messages to the Logo forum to logo-l@gsn.org. Mail
questions about the list administration to logofdn@gsn.org. To
unsubscribe send unsubscribe logo-l to majordomo@gsn.org.