Piligrimus is a neural network indicator. - page 3

 
sab1uk >> :

>> it's always like this... it's all by eye.

God created the world by eye. The trowel and plumb line were invented by the masses to lead everyone astray. (c) Papa Clementius (I think) ;-))

 
Piligrimm писал(а) >>

In this case, I use PolyAnalyst package for my development, which allows me, after training a neural network, to represent the connection between inputs and outputs as a polynomial, this polynomial is the formalized neural network.

This neural network inherently is a polynomial, and how you record it is a matter of personal preference - who gets a pen on paper, who gets a pencil on the wall.

 
Piligrimm писал(а) >>

The whole formula is a union of several networks trained with different parameters and summarized in one polynomial with its own weight coefficients relative to each other.

The entire polynomial is recalculated at every tick, since the history does not change, and only the zero bar quotes change, while the results of recalculation at the zero bar change too. There is no redrawing.

Exactly, it only redraws on the last bar!

These are screenshots of your indicator (sky colour) at consecutive points in time.

That means it re-crosses!

 

Here's an EA to check, try it out, used Piligrimm indicator and a regular MA (21) ;

H4 GBPUSD

from the beginning of the year

Here is a detailed test and the expert

I'm asking you to ignore the big drawdown, trall tp and stop loss were chosen from the fan, no optimization was done!

And remember traders, your opinion is priceless!

Files:
 
kosa писал(а) >>

Here's an EA to check, try it out, used Piligrimm indicator and a regular MA (21) ;

H4 GBPUSD

from the beginning of the year

The archive has a detailed test and the expert

I've used Piligrimm indicator in the archive, I've used a common MA and a Stop Loss from the fan, I haven't run any optimizations!

And remember great traders, your opinion is priceless!

For comparison, replace Pilligrimus with MA and optimize its period, all other parameters will be as they were. Will the better result after optimization be better/different than with Pilligrimus?

 
Integer >> :

For comparison, replace Pilligrimus with MA and optimise its period, all other parameters as they were. Will the better result after optimisation be better/different than with Pilligrimus?

I have a feeling that if you flatten the MA it will be the same..............but the meanings are different .

 
Infinity писал(а) >>

I have a feeling that if you flatten AF you get the same thing..............but the meanings are different.

Any MA is also a neural network - a single layer with linear perceptrons. No matter how you slice it...

 
You can get better results on 2 mediums... What's the trick, comrades?
 
Yes, yes, the middle one is also polynomial...
 
zfs >> :
Yeah, yeah, middle one too polynom...

Yes, we are trying to create artificial intelligence to adapt to the market and show us "what to do ........".

The other day I went to a 3-ruble course on Fibonacci levels and expansions. They suggested working on the "reverse" method - you build a trading system, you see a signal... but you work in the opposite direction ) I don't know what kind of system it is, there's even an indicator for it.... but i don't even think it works on the theory of probability )

By the way, I went for nothing... It was clearly explained how well the levels work since the 80s on a timeframe of years ) and no new innovations)

>> I'd rather give them to someone from the forum to tell me something interesting.)