What to feed to the input of the neural network? Your ideas... - page 77
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
Where does this knowledge come from?
About incorrectly trained data.
The result of "correctness" in the context of earning what ?
What other tc? ))
I will reveal a terrible secret, but in different periods of time and depending on the price structure, it behaves differently. And some people make money in the morning, others in the afternoon, others in the evening, others on Monday.
Every chart is a working one.
Just because someone can't find a TS doesn't mean that you should blame the chart for the noise. And that's exactly what everyone from beginners to masters do: they blame the bare chart for the noise. And its mandatory "processing".
And the justification is simple: it doesn't work.
Other rules, according to which the chart is built.
I will reveal a terrible secret, but in different periods of time and depending on the price structure, it behaves differently. And some people make money in the morning, others in the afternoon, others in the evening, others on Monday.
Everywhere the chart is working.
Just because someone cannot find a TS does not mean that you should blame the chart for the noise. And that's exactly what everyone from beginners to masters do: they blame the bare chart for the noise. And in its obligatory "processing".
And the justification is simple: it doesn't work.
I'll tell you a terrible secret: almost no one makes money. Stop goat-smarting :) you are either capable of understanding what is being said or you are not. Either you communicate on your definitions, but nobody understands you and you are banished to the desert, or you try to understand what others put into it
I'll chew you up (for the last time)
1) One says "it doesn't work"
2) The other says "but because rubbish in is rubbish out."
3) He is asked "how does it follow that there is rubbish in? And if there is non-moosor, will the system make money? "
4) He merges
I hope I am making myself clear. The person is demanding a mandatory action. Allegedly it will "help". But how to do it to "help" - does not know. Therefore, there is no reason to do "it" necessarily.
From here I began to develop the thought and ask questions: what is rubbish, what is not rubbish, what led to the idea and assumption that there is rubbish on the graph. Then the concept of noise, which is used by 99% of traders in the same(!) context, was introduced here like butter: these are the parts of the price chart that prevent you from making money. If there were no noise, they would supposedly make money. There is no other understanding of noise, it is the same: something that prevents a trader or a robot from making money.
And rubbish and noise may differ in definition, but in essence they are the same: an unreasonable appeal to a concept that the majority has invented to justify their failures. Failures of their models and so on.
As a result - they can't define rubbish, and they can't define noise (to which they appeal), except for references to "flat motion". But somehow they quickly forget that "flat systems" make money on flat.
And if you don't understand what I mean, what I want to convey - just forget it. This is just my imho
And if you don't understand what I mean, what I want to convey, just forget it. It's just my opinion
By rubbish in input - rubbish in output we mean that if you don't know what you are training the model for, it will predict unknown things. In terms of patterns.
By noise is meant something that is impossible to predict in principle.
In financial markets, with great uncertainty, it is common to assume that they are random. But you can assume with some probability that the model will "last" for some time on the real. Based on your statistical estimates. That's what preprocessing is for, to know what exactly you are training and in what form. And there will always be noise in the form of errors unaccounted for by the model.
That's the end of my credentials, because I don't even see any problem in understanding these very simple simple truths.
By rubbish in input - rubbish out, it means that if you don't know what you are training the model for, it will predict unknown things. In terms of any patterns.
By noise we mean that it is impossible to predict in principle.
In financial markets, with great uncertainty, it is common to assume that they are random. But you can assume with some probability that the model will "last" for some time on the real. Based on your statistical estimates. That's what preprocessing is for, to know what exactly you are training and in what form. And noise will always be in the form of errors unaccounted for by the model.
That's the end of my credentials, because I don't even see any problem in understanding these very simple simple truths.
But it looks like postulates.
So be it. And let it be that I'm wrong.
But it all looks like postulates .
You can't be right (except by accident) because you are basically a nobody in IT. So it may look any way you like to you, but to people in the field, it's a general and obvious principle :) And they don't even know you exist and didn't intend to hurt your feelings .
And so do I.