Machine learning in trading: theory, models, practice and algo-trading - page 1587
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
I've been talking to him for a while now.
I write about the search for patterns, he throws some kind of stucco in the form of martingale, without explanation.
The main thing in BP prediction is patterns (read: repeating atomic events that can be predicted). If you have anything on the subject, I'll listen and even pretend to sympathize.
The conversation started with this, and then he got stupid for some reason. Some kind of verbal diarrhea.:) Everyone has their own thoughts in their head, I didn't quite get the point about the grids either.
I just noticed the idea that seemed important to me and focus on it. Topic about MO, so I say that the primary strategy - that is, the basic pattern, should be taught separately from the rest of the secondary.
What is required of the primary is SUSTAINABILITY.
And often the profitability of the system as a whole depends to a greater extent on the secondary, but without a stable primary, of course, everything is just an adjustment. Many bump into this at the beginning.
Your findings are studying, not digested everything yet, but I'm trying).
By the way, can you answer-explain
it seems to me that all statistical studies and MO in the field of trading work with returns. Why?
Why? Were there any attempts to apply statistical methods to graphical, candlestick analysis and other higher level stuff?
It seems to me that all statistical research and IOs in the field of trading work with returns.
Have there been attempts to apply statistical methods to charting, candlestick analysis, and other higher-level stuff?
Exactly this is because all MOs work with stationary series, starting from Kolmogorov.
What library are you talking about, fellow citizens? I am writing my own, maybe you can show me a ready one.
And by the way - as it seemed to me all statistical studies and MO in the field of trading work with returns.
Were there any attempts to apply statistical methods to graphical, candlestick analysis and other higher level stuff?
https://www.mql5.com/ru/code/1146 dataanalysis.mqh section
Included in all terminals.
The neural network/perceptron there is slow, so it is better to connect an external one.
But scaffolding and regressions can be used.
And by the way - it seems to me that all statistical studies and MOs in the field of trading work with returns. why?
Returns are the first thing to be analyzed, this is the primary data.
You can also analyze the indicators. But there can be losses and delays, if the indicator is based on MAC.
Have there been any attempts to apply statistical methods to graphical, candlestick analysis and other higher level stuff?
I think someone has tried. I have not.
But wouldn't 50-100-500 consecutive returnees describe any graphical and candlestick patterns?
as the venerable dons believe, is it possible to "cash in" such residual synthetics when it is "conditionally stationary"?
here are a few examples
this is part of the graphs of synthetic residuals
I repeat, the graphs are made at the moment when the synthetic is "conditionally stationary",
the graph starts at the moment of "the beginning of stationarity", it finishes at the moment of "the end of stationarity"
This period can be from 1 to 2000+ bars
how do the venerable dons think it is possible to "cash in" such residual synthetics when it is "conditionally stationary"?
here are a few examples
this is part of the graphs of synthetic residuals
I repeat, the graphs are made at the moment when the synthetic is "conditionally stationary",
the graph starts, at the moment of "beginning of stationarity", ends at the moment of "end of stationarity"
And this period can be from 1 to 2000+ bars
It is not very clear what is meant, but most of the graphs given do not look stationary due to a noticeable trend.
because all MOs work with stationary series, starting from Kolmogorov, the brochure on predictions of stationary series Alexander threw
In our case, one can meaningfully work only with non-stationarity, which in one way or another is reduced to stationarity. Piecewise stationarity, autoregressive models, hmm, etc.
The main reason is that only one realization of the process is always known. For example, if we take speech recognition, there any word we can say as many times as we want. The quotes for a certain instrument over a certain period of time are a single variant. By the way, this is probably the reason why many people here do not distinguish a random process from its realization.
The article is powerful, thank you, if it is not fiction, it confirms the age-old saying about statistics)
Statistics are only a tool. Is it worth scolding a hammer for hitting your fingers?
because all MOs work with stationary series, starting from Kolmogorov, the brochure on predictions of stationary series Alexander threw
But you do not know when your series will stop being stationary
And it may happen right away.
What then? Where do you want the MO to go?
It is not very clear what is meant, but most of the given charts do not look stationary due to a noticeable trend.
Nevertheless, the script displays them as "stationary".
Not without reason it is written everywhere that the series is "stationary" in quotes