Machine learning in trading: theory, models, practice and algo-trading - page 52
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
The fitness may grow simply because the model improved its topology and dropped a couple of neurons, with the same accuracy.
As far as I understood from the comments in the code - it's a port from Lua. And Lua is also a port, because the original was already in c++:http://nn.cs.utexas.edu?neat
Port from port is unfortunate, a lot of unnecessary actions, it would be better if the author of RNeat took C++ code as a basis, and applied the existing R genetic packages to evolve the network.
I found this note on one website, but it's not working anymore for some reason
Maybe someone will be interested :
...The next step was a new model: we apply a low-pass filterto the closing price (I used a 2nd order Butterworth filter), apply a polynomial-harmonic approximation, transform A*cos(wx)+B*sin(wx) to M*sin(wx+f) and take M and f as secondary signs.
.... And with this model I was able to build a network which had very good generalizing properties: the new data almost all correctly recognized.
Yury Reshetov- Yury do you understand what the man wrote here in general?
Something is fashionable to brag about cent accounts lately
mytarmailS:
...The next step was a new model: we apply a low-pass filterto the closing price (I used a 2nd order Butterworth filter), apply a polynomial-harmonic approximation, transform A*cos(wx)+B*sin(wx) to form M*sin(wx+f) and take M and f as secondary features.
.... And with this model I was able to build a network which had very good generalizing properties: the new data almost all correctly recognized.
Yury Reshetov- Yury, do you understand what the man wrote here in general?
To be honest, not really. There are a lot of questions:
It looks like we have some kind of sci-fi text without specifics - unsubstantiated assertions.
And therefore, not knowing the answers to the above questions, we can only guess how to reproduce it in practice and will any adequate result be obtained?
Was there a commercial somewhere again?
The problem here is not bragging or cent accounts.
The problem is something else.
In my opinion, the panelists are hoping to find some kind of grail by slipping them unnecessary inputs.
And they completely forget that they have to use ready-made rules of a strategy or tactics, whichever is closer.
The machine does only what you put in it.
The crap on the input is the same crap on the output.
Excuse me.
The problem here is not bragging or cent accounts.
The problem is something else.
In my opinion, the panelists are hoping to find some kind of grail by slipping them unnecessary inputs.
And they completely forget that they have to use ready-made rules of a strategy or tactics, whichever is closer.
The machine does only what you put in it.
The shit on the input is also the shit on the output.
Excuse me.
A protein neural network can be full of crap at the input and output, too.
Sorry.
A protein neural network can also be full of goo, both input and output.
I'm sorry.
Exactly, that's what I'm saying.
What do you have to say on the subject?
Exactly, that's what I'm talking about.
Do you have anything to say on the subject?
So the question is for you. )
What's the tactic? What's your head got in mind? So tell me.