Machine learning in trading: theory, models, practice and algo-trading - page 1501
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
What's the problem? Make a DLL wrapper, and all the necessary libraries are yours. I do not need to explain it to you.
I understand all this, but in reality it is so that all popular MQL stuff is exclusively in Python, I study Python, but I do not have time for everything, and I also need Python and MQL to make friends - there are ready implementations.... I need time, time and more time (((
With MQL or C++ or C# it's easier, the languages and the logic of coding are 90% similar, but the problem - there is nothing new - everything is in Python, even Microsoft CNTK is a new library, but they do not even provide examples of using for C# - look, there is the same package in Python, just look (((
You may have a look at???? ;)
the question is rhetorical ...
Open Maxim's articles, he writes very briefly and essentially, his work gives good results, he already voiced the problem with this approach - we need a method that determines the need for model re-learning - if you develop it, you'll have a model for any symbol in "2 clicks", ie, the problem of determining the stability of TS in the future ... like everywhere else (((
open Maksim's articles, he writes very briefly and essentially, his works allow to get not bad results, he already voiced the problem with this approach - we need a method that would determine the need for additional model training - if you develop, you will have a model for any symbol in "2 clicks", i.e. the problem of determining stability of TS in the future... well as everywhere (((
Firstly - Do not hide behind Maxim's reference to his articles
Secondly - It turns out that there is no grail after all? And why bring mql into it? )))) You little liar!)
Thirdly - if you solve the problem of the stability of TSin the future, you may trade using a stochastic, knowing what parameters will be profitable in the future
First of all - Don't hide behind Maxim by referring to his articles.
Secondly - It turns out that there is no grail after all, so why write that there is one? And why bring mql into it? )))) You little liar)))
Thirdly - if you solve the problem of stability
1. I'm not hiding, his work read, figured out and checked, very sensible and easy to "finish" the code - I set aside, studied the basics of MO, need a knowledge base to prepare, so that the method of scientific experimentation is not looking for the Holy Grail
2. There is no grail and can not be, but Maxim last year said well that, firstly, it is interesting, and secondly, the self-learning system saves time in finding and checking the TC (I can not guarantee the literalness) .... I am not small!
3. yes, this is the problem, those who really trade, the problem of stability is solved by money management and analyzing the profitability of the TS for trading periods, for MO I would like to think of something cooler, I wrote a couple weeks ago about logit regression, perhaps it can highlight the probability of "guessing the TS" positive expectations in the future
1. I'm not hiding, I read his work, figured it out and checked, very smart and easy to "finish" the code - I put it aside and started studying the basics of MO, I need to prepare the knowledge base, not to seek the Grail by the scientific method
2. There is no grail and can not be, but Maxim last year said well that, firstly, it is interesting, and secondly, the self-learning system saves time in finding and checking the TC (I can not guarantee the literalness) .... I am not small!
3. yes, this is the problem, those who really trade, the problem of stability is solved by money management and analyzing the profitability of the TS for trading periods, for MO I would like to think of something cooler, I wrote a couple weeks ago about logit regression, maybe it can highlight the probability of "guessing the TS" of positive mathematical expectation in the future
The problem is non-stationarity and lack of cycles, basically. By the method of scientific groping I managed to taste what is written about in scientific articles, i.e. "efficiency" of the market. But there are still unexplored areas in the application of MO, maybe later I will formulate and set the task. What is little researched is often there dough, there is no open access to it anywhere, I suppose.
About logit - I sent the article for review. Another mini-study.The problem is non-stationarity and lack of cycles, basically.
Yes, and non-stationarity in everything - in everything!
- I've counted the time between the breaks of the ZigZag - you have to be smart to think up how the market works, there are no temporal repetitions, any combination will not repeat in the near future
- I considered the bar open price on all TFs, the bar height to the high/low, there are no repeats, there will always be a sequence different from the previous one
- I considered patterns with different methods, there is no recurring "plan" of price movement in the future after the appearance of the pattern
.... The non-stationarity and absence of cycles - this is what repeats in the future, i.e. the "history repeats itself" stories - should be read exactly in reverse, what happened yesterday will not repeat today, what happened last month will not repeat this month and so on for all TF
SZZ: imho, the ideal generator of a random sequence is the price ))))
I understand all this, but in reality it is so that all popular MQL stuff is exclusively in Python, I study Python, but I do not have time for everything, and I also need Python and MQL to make friends - there are ready-made implementations.... I need time, time and more time (((
With MQL or C++ or C# it's easier, the languages and the logic of coding are 90% similar, but the problem - there is nothing new - everything is in Python, even Microsoft CNTK is a new library, but they did not even make examples of using for C# - look, there is a similar package in Python
The problem is non-stationarity and lack of cycles, basically
Yes, and non-stationarity in everything - in everything!
In my post I just posed the problem of non-stationarity, in a messy way, but I did
https://www.mql5.com/ru/forum/86386/page1499
Not a single intelligible comment!!! Not to mention an attempt to solve the problem...
In my post I set the task of solving the problem of non-stacynaire, inaccurately, but I did
https://www.mql5.com/ru/forum/86386/page1499
no intelligible comments!!! not to mention an attempt to solve the problem...
how strong a lens should be to see even a hint there?
and in what is the width, height, and speed of the market measured, in what parrots?How strong should I take the lens to see even a hint of it?
and what is the width, height and speed of the market measured in what parrots?The AFC, what else? there is nothing else, all the rest is subjective nonsense
1) A harmonic has only three parameters - amplitude (height), frequency (period) and phase (start point).
2) A function of any complexity can be described by the sum of harmonics
The AFC, what else? there is nothing else, all the rest is subjective nonsense
1) Harmonics have only three parameters - amplitude (height), frequency (period) and phase (start point).
2) A function of any complexity may be described by the sum of hormonics.
This is to Asaulenko and other amateur radio operators