Dependency statistics in quotes (information theory, correlation and other feature selection methods) - page 5
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
The first and extremely vague step is detrending. I have been scolded for posting graphs, but it follows from them that the method used in this thread to get the residual is meaningless, because the statistics of that residual are as bad as the statistics of the original series.
You should read the title of the topic. Are you familiar with the concept of Feature Selection? It is about applying linear and non-linear correlation and other statistical methods to select variables that carry information about the state of zero or future bars.
I try to approach the subject without imposing any subjective limitations, conventions or theories on it
Then go to a kindergarten, where there are no theories, but here people are educated
And if people like you here are educated, I'd advise them to sit behind a desk again.
Comedian
And what prevents from doing it with respect to return? It can be discretised, it is a random variable. Quite a decent object for the information theory application. What identity search can there be? You're playing a war game, my dear...
faa1947: To me TI is about coding and encryption and the reverse process. Here one tries to derive some actually meaningful information on the basis of formulas, supposedly from TI. For such operations, there are other sciences, which besides analysis also prove that we really see what we see, and not some phantom, which only the author of the topic sees.
And for me in this task TI is primarily a tool for datamining. What to do with this data is another matter. The important thing is that we really see something that is not visible to the naked eye. And what other sciences are you talking about?
The article concludes that there are statistical correlations between the increments of quotes. One of these dependences is well-known - the dependence of the conditional variance of the increments (d[t]) on the value of the previous increments (r[t-1], r[t-2], ...) and the variances (d[t-1], ... ).
The reason for finding statistical dependence was most likely the volatility.
Looks like GARCH(p, q). What orders of the model can we talk about?
And how do you see volatility there, if it was only about returns?
By the way, the book on semi-invariants that you advised me may be quite useful for establishing the order of dependence.
No offence to the active participants in the topic. It reminds me of the joke about the plane with all the swimming pools and gyms and stuff. "And now with all this stuff we are going to try to take off". The question is not idle, what is the application aspect? That is, at what point is the transition from theory to practice planned?
Slides, slides... ) This is also from an anecdote.