Market etiquette or good manners in a minefield - page 61
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
paralocus,
You got it?Quite! Where do we start?
What's wrong with the grid in your picture? Did you retrain it?
I'm going to try the same chart now.
I honestly don't understand all this misery. There are so many already written C++ networking codes with ORO on the web. Here, for example, is a simple code with a good description of the theory. There is one bug in calculating MSE (read last post on this page)
http://www.codeproject.com/KB/recipes/BP.aspx
I spent only one day to understand the code and it worked reliably for me in forex (not in the sense of profit, but in the sense of learning the network on currency pairs).
I think you spend a lot of time adapting matcad to neural networks, and then you are going to drag and drop its formulas into C++ and debug the code there. Not efficient, gentlemen. You can spend years like this before you start something that works in forex.
Before you go such a long way with graphical matkadic introduction, C++ extension and forex conclusion, think about what conclusion you are expecting. What advantage would a neural network give you? Get to the bottom of networks before you go on your way. And the bottom line: a network that uses data from the same time series will be an autoregression; linear if you have a single layer, or non-linear if you have more than one layer with non-linear activation of neurons. Training such an autoregressive network is nothing else but an approximation of the series by a non-linear function. That is the difference of such description from the fitting of polynomial or Fourier series is very small. The network's prediction of future values is nothing more than an extrapolation of the fitted non-linear function into the future. The only advantage of a neural network is the universal ability to approximate any non-linear function. The question to ask here is: why do you think that the current price is a non-linear function of past prices? Just because you were able to fit a non-linear function to past prices doesn't mean that you have found a market model. Therefore an autoregressive network will not give you any trading advantage.
...not in terms of profit, but in terms of learning how to network on currency pairs.
Well, we need it in the profit sense...
We might not think so. What makes you think that? We teach hippos to fly and we know they can't fly. Often we don't think anything at all...
Yeah, creating a robust system is just like that, i.e. trying to do the impossible. You get a lot of bumps in the road before you've weeded out enough directions that are unpromising.
Hello, Fedor! I see you'll soon be teaching others how to make a nervous net yourself.
No, I won't teach that. Stupid, with a living master...
I have other things I can teach -:)
Well, we need it in terms of profit...
We might not think so. What makes you think that? We teach hippos to fly and we know they can't fly. Often we don't think anything at all...
Then explain why you think prices are described by non-linear functions.
Here is my explanation. Take the Dow Jones Ind Av as an example. From 2003 to 2007, the index was going up. No network or other non-linear model trained on that data will predict you the crash of 2008. There wasn't enough data on the network inputs to predict that. Stop adjusting your curves to prices. This is child's play. I've posted a bunch of codes myself here for people to have fun with. If a functional model worked in the market, I wouldn't give out these codes for free.
Well then explain why you think prices are described by non-linear functions.
I've already explained to you that I don't think so. I think that sometimes prices can be described by linear functions and sometimes they cannot be described at all (within known mathematics). I want to learn how to be good at predicting what can be predicted and hammer out the rest. My thinking will become much clearer to you if you refer to my very first responses to your posts.