Dependency statistics in quotes (information theory, correlation and other feature selection methods) - page 28
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
Та же возвратность тоже является установленным свойством ценовых рядов
Just because market distributions are of the Parreto-Levy form does not mean that they are returnable. By the same HMT, non-uniformity of inputs determines the clustering of volatility, which in turn entails an increased accumulation of small percentage changes in returns that exceeds the normal distribution. But none of this says anything about returns. There is simply no information (external influence), so there are no trades (the market is in equilibrium), so there is no movement, and the mere absence of movement does not indicate that price is ready to go back.
But all this is irrelevant to the topic.
Just because market distributions are of the Parreto-Levy form does not mean that they are returnable. By the same HMT, non-uniformity of inputs determines the clustering of volatility, which in turn entails an increased accumulation of small percentage changes in returns that exceeds the normal distribution. But none of this says anything about returns. There is simply no information (external influence), so there are no trades (the market is in equilibrium), so there is no movement, and the mere absence of movement does not indicate that the price is ready to go back.
But all this is irrelevant to the topic.
In this case I understood it to mean the market's desire to return to past prices (return).
Returns means return, which again in our sense is the percentage change in price over period t.
Returns - means return or return, which in the context of the market can be interpreted as the desire of the market to return to past prices.
When referring to returns it is better to use the word "return" or to say "returns" and when referring to return it is to say return.
In this case I understood it to mean the market's desire to return to past prices (return).
Returns means return
, which again in our sense is the percentage change in price in period t.Returns means return
or return, which in a market context can be interpreted as the market's desire to return to past prices.When referring to returns we should use "return" or say "returns" and when referring to return we should say return.
Has anyone ever wondered whether price movements can be compared to those of a lift in a high-rise building with heavy inter-floor traffic? Is it possible to predict the position of the lift using the concepts "bar", "TF", "trend", "flat", "levels", "trends", ....?
easily, usually the lights on the 1st floor are on!
In
this case I understood it to mean the desire of the market to return to past prices (return)
.Returns
as return
, which again in our sense is a percentage change of price in period t.Returns
as return
or return, which in market context could be interpreted as a market desire to return to previous prices.When we mean returns we should use "return" or say "returns", but when we mean return we should say return.
It is possible that this is the case. But when we build a returns series of the following form: X[t]-X[t-1], it almost doesn't show it. I use the words returns, increments, returns, they are all a differentiated price series.
The skew of probability in the direction of sign change is minimal and insignificant. But if you calculate the conditional entropy between the dependent variable and returns over two or more lags, then all the unevenness is accounted for in the resulting figure so that entropy is reduced.
I tried to train NS on hourly data and took only the most informative lags (42 variables, on lags 1, 2, 23, 23, 25,... 479, 480, 481). Unfortunately, the result didn't work out very well. Accuracy of prediction of quantile number - in the region of 30-40%. Although, the irregularities the neural network was able to translate to the output, but the dependencies are not sufficient for prediction. The whole problem is that the independent variables are mutually informative at lag 1, 2, 24.... and the total amount of information about the zero bar is really small. We should think as an option to take daily and older timeframes.
Nothing prevents us from checking this information process for stationarity and then applying all econometrics in one fell swoop.
I do not quite understand.
Econometrics works with non-stationary processes, the approximate algorithm is described in the post. We should understand that non-stationarity leads to the fact that we cannot take the best indicator or a set of indicators and get TS and trade stably, because due to non-stationarity any estimates of TS (PF, drawdown and others) are fictitious and in the future there will appear such areas of quotient, where TS will sell out the deposit.
The science of measuring economic data - econometrics, has differences from other very respectable sciences, but it is a separate independent science and proposes to act consistently, fixing each intermediate result as a model, aiming to get a stationary residual, gives stability estimates of future TS when working on a non-stationary market.
This is shown by an example for EURUSD and three indicators (straight line, exponential smoothing, Hodrick-Prescott filter) here.
Guys, let's use a separate science to measure economic data, and not try to pull something out of the neighboring sciences, just because we are too lazy to read the econometric textbook. In our country, there are such textbooks dating back to 2000, i.e. for more than 10 years, universities have been producing specialists who measure economic data scientifically and do not suffer the crap called "information dependence".
And in general, let's live in peace.
Didn't quite get it.
Econometrics works with non-stationary processes, the approximate algorithm is described in the post. We should understand that non-stationarity leads to the fact that we cannot take the best indicator or a set of indicators and get TS and trade stably, because due to non-stationarity any estimates of TS (PF, drawdown and others) are fictitious and in the future there will appear such areas of quotient, where TS will sell out the deposit.
The science of measuring economic data - econometrics, has differences from other very respectable sciences, but it is a separate independent science and proposes to act consistently, fixing each intermediate result as a model, aiming to obtain a stationary residual, gives stability estimates of future TS when working on a non-stationary market.
This is shown by an example for EURUSD and three indicators (straight line, exponential smoothing, Hodrick-Prescott filter) here.
Guys, let's use a separate science to measure economic data, and not try to pull something out of the neighboring sciences, just because we are too lazy to read the econometric textbook. In our country, there are such textbooks from 2000, i.e. for more than 10 years, universities have been producing specialists who measure economic data scientifically and do not suffer the crap called "information dependence".
And in general, let's live in friendship.
I read your article, by the way. It's a valuable article, and the problem of non-stationarity is well addressed there. And I agree that the non-stationarity of financial data is a real and pressing problem. For many months when I was mastering neural networks I tried different transformations of initial time series to improve its stationarity, because NS are sensitive to this phenomenon and learn inadequately. And to be more exact, the error density on the output data is obtained unevenly that in practice leads to strong drawdowns (however with a generally positive MO of the model).
Let's just say for now we've tried it simply on raw data (not quite raw, but a differentiated series) just to see what happens. I'm in no way diminishing the importance of econometrics, although I haven't read any textbooks.
When I have time, I'll post my version of data preprocessing, which, by eye, produces a more stationary series, but I haven't done stationarity tests.