Neural networks, how to master them, where to start? - page 10

New comment
 
The funny thing is that all this mathematics is not very much related to profit. I've seen a couple of TCs, or rather I made them myself, but they predicted the market very well and couldn't make profit. If we are dealing with getting profit, i.e. growth of equity, we should not only take into account results of training, parameters, errors of the net itself but also conditions that form the signal which in its turn may and should be optimized. Maybe I just do not understand what I am doing here. Maybe for you the profit is not the main thing, but the main thing is the best training results.... correct me if what.....
 
Well, that's just me.... to keep the conversation going :)
 
yes, that's the hard truth of life...
Dear budimir, do you consider this third parameter in your work ---> the seasonality factor?
 
the seasonality factor is present if there are cyclic patterns of BP, for that
it is necessary to conduct a preliminary ACF-analysis of BP, and for the first 2 factors
- coefficients from the formula of linear regression (written in mql-language) can be taken from here.
 
budimir писал(а) >>
the seasonality factor is present if there are cyclic patterns of BP, for that
it is necessary to conduct a preliminary ACF-analysis of BP, and for the first 2 factors
- coefficients of the formula linear regression (written in mql-language) can be taken from here.

And ACF can be taken here'Autocorrelation function'.

budimir could you explain on this example (there is a picture), how do you take seasonality out of it (formula if it's not too complicated) ?

 
Prival >> :

And the ACF can be taken here'Autocorrelation function'.

budimir, could you explain by this example (there is a picture there), how do you take seasonality out of it (formula, if it's not too much trouble)?

The point is that I don't perform ACF analysis of BP in mql-language, but I do it

in StatPlus (this add-on in Excel), not to be unfounded, give

screenshot:

As you can see from the figure, in Exel installed add-on StatPlus, in the list SmoothingFactor

this third parameter is not defined ie zero, but it can be calculated with the

ACF-option of this add-in, here's a screenshot of this option:


I'm sorry, but my version of StatPlus is out of date.

 
budimir писал(а) >>
the seasonality factor is present if there are cyclical BP patterns

budimir , in your opinion, is this a promising direction? What percentage of profit can be expected without taking into account brokerage commissions, from the exploitation of the seasonal component on the financial markets?

 
Someone reply to my post on page 9!!!
 
Neutron >> :

budimir, in your opinion, is this a promising direction? What percentage of profit can be expected without taking into account brokerage commissions from exploitation of the seasonal component on the financial markets?

I don't think it's worth taking the seasonal component into account, that's why I nullify the third factor.

 
Andrey4-min писал(а) >>

Am I correct in assuming that the input data means the variables in the external parameters of the EA, with which the coefficients will be compared?

This is how I see the coefficients of a simple Expert Advisor based on fractals:

What should I do with all of them?

The input data for the Expert Advisor (extern ...) are net coefficients, indicator delays Per 1, 2, 3, classification levels u,v, Take Profit and Stop Loss. The indicator (I chose WPR as an example) is calculated inside the Expert Advisor using iWPR.

1...34567891011121314151617...19
New comment