Econometrics: one step ahead forecast - page 28

New comment
 


Here's the prediction algorithm:
Calculating the current currency movement function

constructing a computational trajectory

И.... TOTAL BUMMER ON THE NEWS...

that's it... and now the most accurate, mathematically correct forecast is not worth a penny.

 
new-rena:

What is the average price taken as a basis? I don't see the DX in the terminal.

The formula is not entirely clear to me, as speculation has emerged about the weighting of currencies. How true is this percentage inclusion in the basket of currencies?


I have DX

I propose to do the following. Build a model purely on DX and build a model on the pairs on which DX is calculated. At the same time let's estimate the coefficients you doubt. Will this work?

 
Debugger:

you can make mathematically accurate and calibrated predictions a thousand (1,000) steps ahead...


You can't. The error at 1,000 steps adds up, and it is already large at 1 step.
 
faa1947:
You can't. The error at 1,000 steps adds up, and it is already large at 1 step.


If you know the function there is no error

 
Mathemat:
And what is the point of predicting 4 points with an error of 57 points? I think it's clear without the zero test that zero is being predicted.
Finally. There is a prediction, but there is no estimate of the prediction. This begs the question: Should I take only those that are larger than the spread, stop levelling? Or larger than the forecast error? What to do?
 
This thread is the usual ramblings of amateurs trying to get ideas to solve their problems.
 
faa1947:

I have a DX

I propose to do the following. Build a model purely on DX and build a model on the pairs on which DX is calculated. At the same time let's estimate the coefficients you doubt. Will this work?

No problem... Let's go
 
Debugger:
This thread is the usual ramblings of amateurs trying to get ideas to solve their problems.
We'll have to wait and see if anyone joins in with specifics. In the very beginning offered to fix the model and I will test it. Zero so far. Here's something above.
 
new-rena:
No problem... Let's go

For comparability, we take the same model

Here we have:

EURUSD hp1(-1 to -4) hp1_d(-1) hp1_d(-2)

Here HP is the HP filter of EURUSD. Let us take the filter from DX and predict EURUSD

 
Debugger:
This thread is the usual ramblings of amateurs trying to get ideas to solve their problems.
Come on, I wish I had two years of your time on this subject...
1...212223242526272829303132333435...139
New comment