Using neural networks in trading - page 19
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
No, it isn't.
You will surely succeed.
Just let my words come to your mind one day.
1. The dot on NS has been put on this forum several times. Application of NS is a typical system error of the first kind: solving the wrong problem by the right methods. Where is the problem statement? Where is the verbal, informal description of the marketplace? What is there to learn from the market?
2. In econometrics, the NS is used as one of classification methods. No more than that. And classification is useless for forecasting, i.e. it is impossible to build TS only on the basis of NS.
3. the thing that surprises me the most about NS is something else. NS is obviously not a simple thing with extremely high level of abstraction. It takes years to study and apply. Why is this time not spent studying mainstream stuff called "econometrics"? Example. Let's take two TFs, for example H1 and H4 and calculate the regression for each. Regression at H1 above H4 is long, in the reverse - short. We recalculate it on every new bar, a kind of adaptation. For some sections we obtain a workable TS, and the perspectives for identifying the reasons for losses are obvious, and if we go further in econometrics, the tools to improve this TS are obvious. Studying of regressions together with R will take half a year at most. A well-designed tool just for the market ..... No 100 times NS is discussed and every time we make the mistake described above in point 1.
I never cease to be amazed....
1. Where is the problem statement? Where is the verbal, informal description of the marketplace? What are we trying to find (teach) in the market?
Make money from the market.
Why do we need to write verbally? To create a fuss in the forums?) A mathematical description will be enough.
We teach how to make a profit.
2. In econometrics, NS are used as one of the methods of classification. No more than that. And classification is useless for forecasting, i.e. it is impossible to build TS only on NS.
Example: Recognize a pattern, classify it, make a trade in accordance with the class. Ready-made TS, why not? And all within the capabilities of NS you are talking about.
3) What amazes me the most about NS is something else. NS is clearly not a simple thing with an extremely high level of abstraction. It takes years to study and apply. Why is this time not spent on studying mainstream stuff called "econometrics"?
To draw any categorical conclusions about inability of NS to do something, the same years should be spent. Otherwise it's like "I haven't read the book, but judging by the comments it's rubbish". I have spent, yes, years - found something. Maybe you were unlucky. How much time have you spent on NS? I won't say anything about econometrics, although I have my own opinion about it from the comments.)
Yes, but it's too unstable, because I posted the results. I have not got any further than that. I have some ideas but I can't find them, though the problem is not even in the implementation, but in organizing a normal Walk Forward optimization.
And the echo network wanted to publish, too my hands do not get :) although soon additional motivation will appear, I hope.
TheXpert:
Wanted to publish the echo network, haven't got around to it either :) although more motivation will come soon, hopefully.
Will you leave on your own or do I need to remind you?
Still provoking you to retaliate with rudeness? Not gonna happen.
To make money from the market.
What's with the verbal? Shaking up the forums?) A poorly mathematical description will suffice.
Teach how to make a profit.
Example: Recognize a pattern, classify it, make a trade in accordance with the class. Ready-made TS, why not? And all within the limits of the TS you are talking about.
That to make some categorical conclusions about the inability of TS to do something, you have to spend the same years. Otherwise it is like "I have not read the book, but judging by the comments it is rubbish". I have spent, yes, years - found something. Maybe you were unlucky. How much time have you spent on NS? I won't say anything about econometrics, although I have my own opinion about it from the comments.)
Thank you.
I can't accept the answer. It is not on the merits. Trading on patterns is called "technical analysis". You are trying to find another pattern, nothing more. But you cannot go beyond the limitations of TA.
We take two TFs, for example H1 and H4 and calculate the regression for each. Regression on H1 above H4 is long, on the opposite one is short. We recalculate it on every new bar, a kind of adaptation. For some sections we obtain a workable TS, and the perspectives for identifying the reasons for the losses are obvious, and if we go further in econometrics, the tools for improvement of this TS are obvious. И