Searching for an arbitrary pattern using a neural network - page 2
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
You don't need a neural network to search for a pattern. You can search in an ordinary EA. Put a zigzag on it. To detect the presence of a Head and Shoulders pattern, we need to control 1) the position of extrema relative to each other (higher-lower) and 2) (farther away from the zero bar) in the condition.
It doesn't matter how many bars the pattern lasts, it is enough to control the position of extrema relative to each other vertically and horizontally.
Head and shoulders are just an example. Zigzag is a curve indicator, Gann Swings are better.
Not all patterns can be described by the position of extrema. And there is one big mistake in your method: the mutual position of extrema may look like head and shoulders, but it will not be the same in most cases.
Colleagues, think about it, if you cannot formally describe the pattern, how do you plan to explain it to the NS? Based on what? Interesting to hear....
Understand the use of NS implies two things. You can find a pattern mathematically, but you can find rubbish in place of the pattern. So you use NS to separate the rubbish from the pattern you're looking for.
And the second you're asking to find the NS pattern leading to a profit, and in the learning process, it finds it BUT what kind of pattern it is you'll never know because it will be a purely personal pattern of this particular NS trained at this interval at a certain point in time (conditionally).
You should decide what you want to get?
BUT if you want to teach the NS to look for a specific pattern, you have to explain it by mathematics and if you can properly describe 100% of your pattern with mathematics, the NS will not be needed. NS becomes necessary when mathematical tricks are not enough and at some point you need an ambiguous conclusion, welcome to the world of artificial intelligence :-)
Colleagues, think about it, if you cannot formally describe the pattern, how do you plan to explain it to the NS? Based on what? Interesting to hear....
Understand the use of NS implies two things. You can find a pattern mathematically, but you can find rubbish in place of the pattern. So you use NS to separate the rubbish from the pattern you're looking for.
And the second you're asking to find the NS pattern, leading to a profit, and in the learning process, it finds it BUT what kind of pattern you will never know because it will be a purely personal pattern of this particular NS trained at this interval at a certain point in time (conditionally).
You should decide what you want to get?
BUT if you want to teach the NS to look for a specific pattern, you have to explain it by mathematics and if you can properly describe 100% of your pattern with mathematics, the NS will not be needed. NS becomes necessary when mathematical tricks are not enough and at some point you need an ambiguous conclusion, welcome to the world of artificial intelligence :-)
Misha, they have only one error here, they understand a pattern not as a current state of the system, which should be recognized, but only as a type, head-shoulders, three soldiers and a crow on the pole.
Misha, they have only one mistake here, they do not understand the pattern as the current state of the system, which needs to be recognised, but solely as a head-shoulders, three soldiers and a crow on a pole.
In order to recognize the current state of the system, it has to be assigned to a class.
Colleagues, think about it, if you cannot formally describe the pattern, how do you plan to explain it to the NS? Based on what? Interesting to hear....
Understand the use of NS implies two things. You can find a pattern mathematically, but you can find rubbish in place of the pattern. So you use NS to separate the rubbish from the pattern you're looking for.
And the second you're asking to find the NS pattern leading to a profit, and in the learning process, it finds it BUT what kind of pattern it is you'll never know because it will be a purely personal pattern of this particular NS trained at this interval at a certain point in time (conditionally).
You should decide what you want to get?
BUT if you want to teach the NS to look for a specific pattern, you have to explain it by mathematics and if you can properly describe 100% of your pattern with mathematics, the NS will not be needed. NS becomes useful when mathematical methods are not enough and at some moment you need an ambiguous output, welcome to the world of artificial intelligence :-)
And why formally describe a pattern for a neural network? The network must be trained by examples.
Why would a neural network need a formal description of a pattern? A network must be trained by examples.
So you make it first find this pattern and then recognize it? Isn't it too much? Moreover, any pattern head and shoulders, three soldiers, any pattern must be recognized not who it is and what it is called, but what will be the market reaction up or down on it. ALL known patterns mechanically work 50%. And what can be done with mathematics is the best, and when you run into uncertainty, that's when the grid as a mathematical tool.....
So you make her find this pattern first and then you make her recognise it? Isn't it too much? Especially any pattern head and shoulders, three soldiers, any pattern must be recognized not who it is and what it is called, but what will be the market reaction to it up or down. ALL known patterns mechanically work 50%. And what can be done by mathematics is best done by it, and when you get stuck in uncertainty, that's when the grid as a mathematical tool .....
And how did you manage to guess what I might have meant by that?
But if you want to teach NS to look for a particular pattern, you have to explain it by mathematics and if you can describe 100% of your pattern with mathematics, NS becomes unnecessary. NS becomes necessary when mathematical methods are not enough and at some moment you need an ambiguous conclusion, welcome to the world of artificial intelligence :-)
The NS is designed for cases where you cannot describe the situation mathematically. Try, for example, to describe a person's age mathematically from a photo.
Excuse me, but does the neural network read coffee grounds? The way I see it, it does the math.
Excuse me, but does the neural network read coffee grounds? The way I understand it, it does the math.
Do you think there is a difference between "a human describes" and "a neural network counts"?
A neural network is trained by examples, without going into details of dependencies.