Testing real-time forecasting systems - page 26
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
>> what's that?
This is an analogue of the Matcadian Predict function.
Yep, that's the best you can do in MT:
only mountains are cooler :o) Thought they'd all be multicoloured, thick, thin, uninterrupted, beautiful... but it's OK too, it's simple, austere and tasteful. :о))
Well, for fun, here are my predictions of the 4 methods. 3 of these 4 point to a EURUSD move downwards.
1. The nearest neighbour method
2. GRNN
Non-linear autoregressive model of the 4th order (like linear prediction, but with non-linear terms x[i]*x[j], x[i]*x[j]*x[k], ...)
4. Fourier extrapolation
3. Нелинейная авторегрессионная модель 4-го порядка
Where can I read about this method?
And here is my own indicator showing when to buy and sell. The last signal was a sell signal. So expect EURUSD to move down
Where can I read about this method?
See here
https://forum.mql4.com/ru/24614/page3#190545
and on previous pages where I describe shortcomings of neural network mathematics in terms of training (a touchy subject :-)
I could not find suitable links on the web, although the idea is simple and should be familiar to econometricians dealing with linear prediction. Maybe try the Nonlinear Autoregressive Model.
See here
https://forum.mql4.com/ru/24614/page3#190545
and on the previous pages of that thread, where I describe the shortcomings of neural network mathematics in terms of training (ticklish topic :-)
I couldn't find suitable links on the web, although the idea is simple and should be familiar to econometricians dealing with linear prediction. Maybe try Nonlinear Autoregressive Model.
Got it, thanks.
As for my prediction, the central curves are essentially the 'average' to which the future trajectory should statistically 'gravitate'. What it will actually be, we shall see.
Well, for fun, here are my predictions of the 4 methods. 3 of these 4 point to a EURUSD move down.
1. The nearest neighbour method
2. THE GRNN
Looks like GRNN has hit the jackpot :-)
Here's what my prediction would be (no adjustable parameters, no fitting)
Looks like GRNN has hit the jackpot :-)
Here's what my prediction would be (no adjustable parameters, no fitting)
no no no,
here I'm posting the forecast at the beginning, not showing it at the end
also, if there are no parameters, how does gpwr get a different line?
No, no, no,
I'm posting a forecast at the beginning, not showing it at the end.
besides, if there are no parameters, how does gpwr get another line?
Sorry :-) Here is the forecast at the very beginning:
gpwr apparently prepares the Fourier differently, my method has no parameters.