Econometrics: State-space model forecasting - page 12
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
On the topic of the thread, I've already told you.
If "predicts well", then trade in the direction of the prediction.
If such a trade leads to losing on the spread, then the forecasts are bad.
You can check the "goodness" of predictions in the strategy tester.
EconModel:
What a bicycle: it's about applying the listed packages, nothing more, not enough brains for that
That's right, you have to work out the methodology first.
That's right, you have to work through the manual first.
On the contrary.
Well, so do I. I can assure you that the documentation on state-space models is far from simple.
In some universities, the state-space models are studied for three years at a time. This level of knowledge cannot be achieved by reading the documentation.
In some universities, the state-space models are studied for three years at a time. This level of knowledge cannot be achieved by reading the documentation.
You are absolutely right. The documentation assumes prior knowledge.
I spent six months trying to build a model in state-space, and then I discovered that I had passed the required modification as an exam. That's what I'm using now. I didn't understand how to use the point forecast, thanks to faa, although I haven't run it in the tester yet, it looks very similar to the truth.
The model is made in R. So there is a bunch of related information. Here is the MSE - Mean Square Error - Mean Square Error. This is the square of the error. Below is the graph, but an error comparable to the eurusd level can be obtained by extracting the root.
I.e. for the maximum error the comparable one is 0.001871
Now I understand: it is necessary to enter, if the forecast exceeds that 18 pips? Or the average? Or the RMS?
The RMS is only and exclusively good for a normal distribution.
The real distribution is very different from the normal distribution. It has much thicker tails. And hence the risks are much greater than those estimated by the normal model.
The RMS is only and exclusively good for a normal distribution.
The real distribution is very different from the normal distribution. It has much thicker tails. And hence the risks are much higher than those estimated by the normal model.
About the error, it is assumed to be iid. But all tests are unit root tests (the result was given). It is very desirable to look at ACF. I've never seen an error (residual from model) tested for normality. The reason for this is not known to me.
The problem of biased estimates is solved by bootstrapping and resampling.