Dependency statistics in quotes (information theory, correlation and other feature selection methods) - page 4
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
2 faa1947: OK, let science. As I recall, econometrics is very fond of imposing its models on financial data. I don't impose them. So I'm not doing econometrics. Any other questions?
Stop and put yourself on EViews
I already have it up and running. But I haven't started on it properly yet. It's nice to hear familiar names...
With the word "moderator" in mind, I want to complain about life. In my estimation the whole thing (with the rarest of exceptions) is a song of a Chukchi who drives across the tundra and sings - what he sees.
Econometrics is a pretty old lady and has a huge amount of tools that you still need to know how to apply. I am not an econometrician either, our universities have been graduating these specialists for more than a decade, but for some reason they are not on this forum.
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 derive the residual is meaningless, as the statistics of this residual are as bad as the statistics of the original series.
I have been trying to publish a starter paper on the application of econometrics for almost 2 months, but so far without success. Maybe with its publication it will be possible to organise a coherent discussion on BP analysis and the synthesis of models to make predictions with its error.
And EViews is considered one of the best econometric programmes. And the level of training of our specialists is probably not "mature" enough - if you are talking about quantitative finance. I think they will make themselves known very soon.
The first and extremely vague step is detrending. I have been scolded for posting graphs, but from them it follows that the method applied in this thread to derive the residual is meaningless, as the statistics of that residual are as bad as the statistics of the original series.
Yes, vague. Once again: it is a preprocessing method used for many econometric models, for which it is desirable to eventually obtain I(0). Who says it's I(0) that's needed for the branch's research topic?
I have been trying to publish a starter paper on the application of econometrics for almost 2 months, but so far without success. Maybe with its publication it will be possible to organize a coherent discussion on BP analysis and synthesis of models to make predictions with its error.
It's a great idea. I support it with all my heart. What prevents you from publishing it?
2 HideYourRichess: I'm having a bit of a holiday today, so I'm temporarily free to say whatever I think :) Are we having a religious showdown about what information is?
Mathemat:
Who says that the branch's research topic needs I(0) specifically?
Is the research topic here? 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. There are other sciences for such operations, which besides analysis also prove that we really see what we see and not some phantom that only the author of the topic sees. I, on the other hand, gave an example in the post that it was possible to achieve a constant volatility of the residual at all, which tells us that if there was one, it is not significant and should not be dealt with.
This is a great idea. I support it with all my heart. What prevents you from publishing it?
On MQL5 in the "ready to be published" state, it's just stretching and my post is more of an advertisement and an invitation for discussion.
Can you explain on your fingers why this has to be taken into account at a stage where no models of incremental behaviour are being built? No models, just a blunt application of information theory. Thank you.
The article concludes that there are statistical dependencies between quotes increments. One of these dependences is well-known - the dependence of the conditional variance of increments (d[t]) on the value of previous increments (r[t-1], r[t-2], ...) and variances (d[t-1], ...).
The reason for finding a statistical relationship was most likely the volatility. And volatility prediction is unlikely to help in the direction prediction task that alexeymosc is interested in.
I don't understand what you are doing here. I decided to refresh my understanding of the Information Theory (IT) and looked it up in the glossary of terms:
TI considers the notion of "information" only from the quantitative side, without reference to its value or even meaning. With such an approach a page of typewritten text at most always contains approximately the same amount of information, determined only by the number of characters and spaces (i.e. characters) on the page and not depending on what is printed on it, including the case of meaningless, chaotic set of characters. For modeling communication systems, this approach is valid, as they are designed to transmit information represented by any character set error-free over the communication channel. In those cases where it is essential to consider the value and meaning of information, the quantitative approach is inapplicable. This circumstance imposes essential restrictions on the fields of possible applications of TC. Failure to take it into account led at early stages of development to overestimation of applied significance.
In this connection I have three possible answers:
1. You are sure that the dictionary is lying and that's not really the case.
2. You are in the "early stages of development" and have not yet undertaken an assessment of "applied relevance".
3. You are something else.
So you questioned the very legitimacy of this approach in my article?
Reading your assumptions, we are something else. I cannot say that I am in the early stages, however, I strive to approach the subject without imposing any subjective limitations, conventions or theories on it. The study started with a clean slate, meaning that all sorts of economic and other senses were not applied in interpreting the process. Therefore, I believe that the application of TI formulas is at least not wrong for such a task.
The article concludes that there are statistical dependencies between quotes increments. One of these dependences is well-known - the dependence of the conditional variance of increments (d[t]) on the value of previous increments (r[t-1], r[t-2], ...) and variances (d[t-1], ...).
The reason for finding the statistical correlation was probably the volatility. And volatility prediction is unlikely to help in the task of predicting direction, which alexeymosc is interested in.
The dependence of the variance on past values of the variance could indeed occur. I agree. But, we need to move on, try to remove this dependence and see if it stays.
And I have a counter question, has the applicability of the axiomatics of the term information that TI operates with respect to the market been established?
Will you take up this challenge or will we argue over nothing?