|
PureBytes Links
Trading Reference Links
|
Clearly a NN will be accurate to within a certain
percentage of error. The amount of error depends
on the amount of reliable structure within price
movements. That of course assumes that the NN
can seperate randomness from structure.
And the "prediction" needn't be an exact image
of the price movements in the future, rather it
could be a "probability" of price direction and
movement. This way we aren't saddled with the
paradox of being able to predict the weather
or decrypt intelligable messages from seeming
giberish.
A neural network based S&P 500 trading system that
I developed back in 1994 did just that within an
accuracy that varied from 60 - 80 percent in tests
against historical pricing data.
- John
__________________________________________________________
John T. Nelson | John's Trading Journal
trader@xxxxxxxxxxxxxxx | http://trader.computation.org/
__________________________________________________________
Hosted by Lexonia Internet Services
http://www.lexonia.net/
|