Econometrics: bibliography - page 8

New comment
 
alexeymosc:
Read Ilya Prigozhin. You will learn a lot. Chaos is in all dynamic systems.
Chaos is one of the methods, not better or worse, just trendy. We are talking about something else. About a systematic approach to the construction of TS.
 
faa1947:
Chaos is one method, no better or worse, just trendy. It's about something else. It's about a systematic approach to the construction of TC.
You can also take a systematic approach to chaos. I've read one or two good publications on the application of chaos to stock market trading. It's a bit complicated for me, to be honest...
 
alexeymosc:
You can also approach chaos in a systematic way. I have read one or two good publications on the application of chaos to stock market trading. It's a bit complicated for me, to be honest...
There are many models: regression of various kinds, ARIMA, etc. Some models have no advantage over others. Chaos stands apart and there are so many issues that have yet to be resolved when applying it.And the problem is not that it's hard to figure out. It's just a fashionable fad.
 
alexeymosc:
You can also approach chaos in a systematic way. I have read one or two good publications on the application of chaos to stock market trading. It's a bit complicated for me, to be honest...
We've been there. They change a couple of parameters in the formula of standard deviation - and proudly announce: "This is Chaos!!! Tremble! Our new method is based on the chaos theory!" But in fact, they take an ancient statistical apparatus chewed up a hundred times over, wrap it up in a beautiful wrapper "CHAOS" and present it as a new dish. But you cannot cheat reality. Let us read the preamble of the book and trash it. If the preamble (idea) makes sense, develop a new methodology for it.
 
C-4:
We know, we've been there. They change a couple of parameters in the standard deviation formula - and proudly announce: "It's Chaos!!! "It's Chaos!!! Our new method is based on the chaos theory!" But in fact, they take an ancient statistical apparatus chewed up a hundred times over, wrap it up in a beautiful wrapper "CHAOS" and present it as a new dish. But you cannot cheat reality. We read the preamble of the book and trash it. If the preamble (idea) makes sense, develop a new methodology for it.
) In fact, they take quotations of American stocks (there are thousands of them, you can choose...), daily quotes. Then we look for the dip parameter in the lag space, calculate the Lyapunov exponent and other stuff. But to be honest, it does not give such a nice result...
 
Why do I stick with EViews? Because you open it up and start implementing an idea (Lyapunov) and you always see how much more needs to be done. Everything and more is in Matlab, but there you don't see how much you haven't done. At the same time you realise that if you have done everything that EV suggests, there is no guarantee that some bad thing will come out and drain the deposit.
 
alexeymosc:
) Basically, it takes US stock quotes (there are thousands of them, you can choose...), diaries. Then we look for a dip parameter in the lag space, calculate Lyapunov exponent and other delights. But frankly speaking, it doesn't yield such a nice result...

Once again. You can search in the lag space even dark matter (there is such an activist here), but if your calculation is devoid of common sense, it is all the same what Lyapunov exponent and where you will apply it. It will not give the result, as you yourself say.
 
C-4:

Once again. You can search the lag space even dark matter (we have such an activist), but if your calculation is devoid of common sense, it does not matter which Lyapunov exponent and where you will apply. It will not give the result, as you yourself say.

Yes. I agree. You could consider the correlation between cow yield and the degree of baldness of the milkmen. ) The point would be lost, although "the algorithm works".

 

So has anyone actually calculated this very Lyapunov exponent on forex?

in the sense of multi-currency analysis...

I think the standard correlation coefficients are not very suitable due to the non-linearity and lag of the series.

 
avatara:

So has anyone actually calculated this very Lyapunov exponent on forex?

in the sense of multi-currency analysis...

I think the standard correlation coefficients are not very suitable due to the non-linearity and lag of the series.

We consider cointegration, Glanger causality test, and correlation for self-deception, for the amateur.
123456789101112131415...19
New comment