Machine learning in trading: theory, models, practice and algo-trading - page 2490
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
That's why I'm talking about the smile, and I still can't do it.
I'll leave it for history -
VBA Break-even volatility skew computation- (taking into account different dte)
- in VBA without libraries is somehow not very convenient sometimes
p.s.
Python (for example, for time series analysis) looks more attractive - thanks to author of article on the link... Thanks to the author of the article with the link... and for the Objective mentioned in his other article:
Let's write a simple Expert Advisor that will use the discovered pattern.
I didn't understand anything, but the music is still humming in my head)
Include subtitles - made with your criticism in mind! Better now?
Include subtitles - made with your criticism in mind! Better now?
I'm not criticizing you at all, just thinking out loud.
I did not understand anything because I did not go into it, and the article glanced through, so that I do not understand - this is my problem, so be calm ))
===================
I understand the deep gist of the approach...
1) We forecast not everything, but choose only some cases from the history, which fit our rules, it can be called "the initial rule / rules of the TS and so on ...".
2) trains the model only on those data when the "initial rule/TC rule" worked out
Thus, we sharply compress/clean the information.
I'll leave it for history -
VBA Break-even volatility skew computation- (taking into account different dte)
- in VBA without libraries is somehow not very convenient sometimes
p.s.
Python (for example, for time series analysis) looks more attractive - thanks to author of article on the link... and for the Target he mentioned in his other article:
I'm not criticizing you, just thinking out loud.
I didn't understand anything, because I didn't go deep into it, and I have glanced through the article, so the fact that I didn't understand anything - is purely my problem, so don't worry ))
Anyway, if you need tutorial on how to run and use CatBoost in MT5 without R and python, read the article and watch the video.
If it is clear, then I don't need it and if not - I'll write what I don't understand.
If you need tutorial on how to run and use CatBoost in MT5 without R and python, read the article and watch the video.
If it is clear, then no one needs it, and if not, then I'll write better what is not clear.
Well, I certainly do not need, I'm doing it in two lines in my R-ka, your simplification for me a pain, and for you on the contrary, since you do not want to work with R-ka, the point is simple, it makes no difference what to do, the important thing is that do, and then if I do not need it just does not mean that others do not need, then I am here in the deepest minority, so do what you think necessary!
The deep essence of the approach I understand...
1) We don't predict everything, but choose only some cases from the history that fit our rules, it can be called "the initial rule / rules of the TS, etc..."
2) trains the model only on those data when the "initial rule/TC rule" worked out
Thus, we compress/clean the information.
In general yes, now I'm developing the concept of "significant event", according to which the market consists of significant events that can change the market trend and at the same time are the points with prediction advantage, respectively there is a time of influence and mutual influence.
Well, I do not need, I do it in 2 lines in my own R-ka, your simplification for me a pain, and for you on the contrary, as you do not want to work with R-ka, it's your business, it does not matter what to do, it is important what to do, and then if I do not need it does not mean that others do not need, I am here in the deepest minority, so do what you think necessary!
Not a simplification, but an alternative that will work faster, and that allows you to batch many computers to find a solution!
Yes I'm about the benefit globally to others, to popularize it so to speak.
In general, yes, now I develop the concept of "significant event", according to which the market consists of significant events that can change the trend in the market and at the same time are points where there is an advantage in the quality of forecasting, respectively, there is a time of influence and mutual influence.
Well, the same thing is done by Misha, his "ground rule" is the "TS Sequence", from which he trains his AMO.
You are essentially doing the same (absolutely) just call things by different names.
But your approach is smarter because you can choose any "starting rule" and he's tied to one...
I'm doing exactly the same, especially if I want to load say 100 million predictors into my AMO. It's just technically impossible without compression/setting rule/TC rule.
For this I invented a "knowledge base" so that not to keep in one table 100 million predictors AMO itself will call the right ones from the right folder at the right time, but it's all still at the stage of reflection...Well, the same does our Misha, his "initial rule" is "TC sequent" from it and he trains his AMO
You are doing one and the same thing (absolutely), you just call things by different names.