Need Neural Network Solution
- Thread starter bgerdsen
- Start date
Aren't we all?
http://home.flash.net/~hesler/picks.htm
http://www.tradingsolutions.com/
http://www.alyuda.com/
On the software side I think you are going to find that most packages require you to select the inputs to the neural net and once you know that (what inputs are correlated with future prices) you don't really need the neural net at all.
Winter,
I agree with your last statement.
As Semper was looking for a 'turnkey' solution, I don't think that selecting inputs to the net and defining net topology is what he would go for.
He should be warned that many 'solutions' have appeared over the last 5+ years and have vanished again in the mist.
try www.neuroshell.com
but it's no turnkey, sorry forgotten the link for that
anyway it's a rock solid program with a very intuitive interface especially for intermarket solutions.
I can also recommend the adaptive turboprop addon, it will let your create predictions that can reoptimize every bar in RT automatically. Though it is generally more interesting to fish for patterns with NN
my best,
cosmic
but it's no turnkey, sorry forgotten the link for that
anyway it's a rock solid program with a very intuitive interface especially for intermarket solutions.
I can also recommend the adaptive turboprop addon, it will let your create predictions that can reoptimize every bar in RT automatically. Though it is generally more interesting to fish for patterns with NN
my best,
cosmic
Most people see a neural network as a miraculous solution.
They think you only have to push a button and wait a few days until all the calculations are done and the correct answer will be given.
That's completely wrong.
A neural network is a tool. If you don't know how to use it it's useless.
I have a neural network. 450 pages of the manuals were filled with how you have to prepare the correct sets of data. The whole system stays or falls with the data you put in. 90% of the people on this forum are probably not able to put the right set of data in the computer to have a change of getting something useful out of it.
... neural network buffs have an unshakable belief in the ability of neural networks to solve most any kind of problem. Undoubtedly, neural networks are one of the most powerful tools ever as far as trading the market is concerned. However, they should be applied at the right time and place, and, for this very reason, can only be used when combined with technical analysis and other things.
PleaseContactBaronToBecomeASponsor.com
=================
To speak the truth, in any event, provided the technology is applied correctly, you will gain more than you will lose, but you will be a real success with neural nets only when you stop looking for the best and, naturally, non-existing net.
PleaseContactBaronToBecomeASponsor.com
PleaseContactBaronToBecomeASponsor.com
=================
To speak the truth, in any event, provided the technology is applied correctly, you will gain more than you will lose, but you will be a real success with neural nets only when you stop looking for the best and, naturally, non-existing net.
PleaseContactBaronToBecomeASponsor.com
You bring up a key point. This is the "magic" behind any and all successfully implemented learning algorithm - getting the data into a format that makes it easy for the technique to tease out meaningful patterns.
I remember when I first threw raw stock data at a neural network back in the 80's. It took me a while to figure out some interesting derivations of the data to make the network at least train. It still didn't do a good job of forecasting, but it was a great learning experience. I really got good at using Lotus123.
Try also http://avis-c.gluk.org
It a new kind of NN.
It a new kind of NN.
http://www-ra.informatik.uni-tuebingen.de/SNNS/
Best open source NN out there in my opinion.
If you cannot put together the data to drive it or cannot go through the documentation the chances of getting any kind of positive result is less than zero. (IMHO)
However, if you just gotta play with NNs without paying a $1000 it is very good.
Good luck,
Best open source NN out there in my opinion.
If you cannot put together the data to drive it or cannot go through the documentation the chances of getting any kind of positive result is less than zero. (IMHO)
However, if you just gotta play with NNs without paying a $1000 it is very good.
Good luck,
