Bayesian regression - Has anyone made an EA using this algorithm? - page 18

 
Dmitry Fedoseev:
Knowing what an arima is and making one are two different levels of competence, that's not what you're asking. And what the hell is she...
 
Dmitry Fedoseev:
Looking forward to seeing the ARIM code in the codebase.

Eat, really AR, but you can substitute anything

I've got plenty of this stuff, but I'm not allowed to use third-party DLLs.

 
Комбинатор:
Knowing what is ARIMA and making it are two different levels of competence, that's not what you're asking. And what the hell is it for?

It seems to me that it is possible to predict the next bar when arbitrating. But I don't think I've tried it.

Let's work together.

I will correct the indicator for ARIMA and you make an Expert Advisor that will trade by predicting the next bar deviation.

And we will see what happens.

 
СанСаныч Фоменко:

Eat, really AR, but you can substitute anything

I've got plenty of that, but I'm not allowed to use third-party DLLs

Without R, purely on mql.
 
СанСаныч Фоменко:

It seems to me that it is possible to predict the next bar in arbitrage.

When you arbitrage, you use cointegration, so you don't have to make up any models to trade.

Thanks for the offer, but I'll pass.

 
Dmitry Fedoseev:
Without R, purely on mql.

Are you suggesting that I swap a sea of professional software for a sub-product, even if it is from me personally?

What are you confused about?

EA template using R has been known for many years...

We send to R another batch of cotier, back the prediction...

 
СанСаныч Фоменко:

Are you suggesting that I swap a sea of professional software for a sub-product, even if it is from me personally?

What are you confused about?

EA template using R has been known for many years...

We send to R another batch of cotier, back the prediction...

If we consider the question of professionalism, then purely on mql will be more professional, if done properly, of course. But, of course, all felt-tip pens are different for taste and colour.

I'm disturbed when everything is not clear in the algorithm, and when there are a lot of extraneous factors. There is no desire to put myself R and work it out.

 
Dmitry Fedoseev:

If you consider the question of professionalism, then purely on mql will be more professional, if done correctly, of course. But, of course, all colour tastes differ.

I'm disturbed when not everything is clear in the algorithm, and when there are a lot of extraneous factors. There is no desire to put myself R and work it out.

No judgement is necessary.

PS.

By the way, are you well versed in the structure of processors you use? different transistors, adders.....

 
СанСаныч Фоменко:

There's no way around it.

PS.

By the way, do you have a good understanding of the processors you use? various transistors, adders.....

One of my favourite children's books was called 'Transistor' and the other was 'Microcircuits and Applications'. There was another one called The Adventures of Pinocchio, so here I am.
 

ARMA-type models can be easily and simply implemented -- provided, of course, you understand the subject matter -- by mql means, without involving anything R-"professional".

zy.

The abbreviation ARMA is AR (autoregressive) and MA(moving average). In ARIMA there is addition I (integration).