Econometrics: one step ahead forecast - page 82
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
???????
Exactly?
Topekstarter: Try to take the first differences of your price series, mix them up, integrate them, estimate the parameters of the proposed model and calculate its coefficient of determination.
What is the goal?
To integrate. To take the difference?
The model does not work for the differences. You can see that in the table above. You got a negative R-square there
What is the goal?
Actually this is the simplest way to check if your model actually works. If there is a much smaller R^2 on the integrated series of mixed price increments, then there really is something in your model.
To integrate. To take the difference?
The model doesn't work for differences. You can see it in the table above. You got negative R-squared there
Read carefully. I didn't suggest applying it to differences.
This is actually the simplest way to check if your model actually works. If there is a much smaller R^2 on the integrated series of mixed price increments, then there really is something in your model.
What are increments and what are intermixed? if you can give an example. Is it relevant to bootstrap?
The increments are returns.
returns(0) = Close[0]-Close[1] in MT4.
Integrated is the accumulated. If we know the initial price on bar 10 and returns from that bar onwards to zero, we can easily find the price on zero by summing all returns and adding the price on bar 10. Here summation = integration.
I don't believe an econometrician doesn't know what increments are.
Bootstrap is different, and has to do with new statmethods with accelerated convergence to marginal distributions.
The increments are returns.
returns(0) = Close[0]-Close[1] in MT4.
Integrated is the accumulated. If we know the initial price on bar 10 and returns from that bar onwards to zero, we can easily find the price on zero by summing all returns and adding the price on bar 10. Here summation = integration.
I don't believe an econometrician doesn't know what increments are.
Bootstrap is something else entirely, and has to do with new stat methods with accelerated convergence to marginal distributions.
ARIMA = ARPSS(p,d,q) is an autoregression of the integrating moving average. d is the order of magnitude of the difference, called the cointegrated. Clarification is still desirable .
The idea is new to me and if I understand it, I will definitely try it.
What are incremental and what are reintegrated? if you can give an example.
Let p[i], i=1...n be a vector that contains the original time series (price values over some period).
1. Calculate the increments of price: r[i]=p[i+1]-p[i], i=1...(n-1)
2. Mix the vector of price increments and get: r2[i], i=1...(n-1)
3. Calculate the cumulative sum of the vector r2: p2[1]=0; p2[i]=p2[i-1]+r2[i-1], i=2..n
Test the model on the obtained data p2[].
Numerical example:
p={0.9379413 0.1411467 0.2540312 1.5440039 1.2363895} // some price series
r={-0.7967946 0.1128845 1.2899727 -0.3076144} // differentiate
r2={-0.7967946 -0.3076144 0.1128845 1.2899727} // shuffle
p2={0 -0.7967946 -1.1044090 -0.9915245 0.2984482} // integrate
Do you have any idea what you are writing, colleague?
I haven't understood anything for a long time. Just bringing to your attention the available terminology, which was invented to confuse the class adversary who does not want to read books.