From theory to practice - page 1705
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
Here is an interesting article - Methodology for determining the optimal sample size for predicting non-stationary time series".
It also makes some kind of forecast of the series
In fact, such regions lead to the very overlap in the histogram of the increments. In terms of the Glivenko-Kantelli theorem there is a violation of the sampling independence condition. Positive correlation (positive correlation coefficient), obviously, leads to histogram peakedness, and negative correlation leads to flatness. In addition to this, sampling heterogeneity (violation of another condition of this theorem) may well lead to a double-top histogram.
For example it is possible, more or less, to remove diurnal correlations from the sample. It's quite simple - everything that falls in the channel around the daily SMA shifted by half a day is discarded :-) Also for 2 and 3 days.
The rest of the channel does not contain correlations of multiples of 24 hours. The only question is how to perform such a channel correctly (obviously, it is necessary to take into account the deviation in width, but not bbands), so as not to throw out too much.
you will have two distributions - the ones you have dropped and the ones you have left :-)
Here is an interesting article - Methodology for determining the optimal sample size for predicting non-stationary time series".
It also does some sort of forecasting of the series.
At first glance, it looks like a return to the original branch idea. That's when we take a sample of one hundred and five hundred trillion increments and use them to draw a histogram, and then we are surprised that the sample of increments for some hours (the time of transaction holding) doesn't correspond to this histogram.
Nevertheless, the article introduces statistics which can be quite calculated and we can think if this non-stationarity model (epsilon-stationarity) is suitable for our purposes. To do so,the minimum samplesize T calculated as suggested there must be comparable to the lifetime of the trades.
e.g. it is possible, more or less, to remove diurnal correlations from the sample. Quite simply, everything in the channel around the daily SMA shifted by half a day is discarded :-) Also for 2 and 3 days.
The rest of the channel does not contain correlations of multiples of 24 hours. The only question is how to correctly draw such a channel (apparently, it should take into account the deviation in width, but not bbands), in order not to throw away too much.
I'm afraid that the algorithm usually works fine separating wolves and goats from the history only.)
I'm afraid we will get an algorithm that splendidly separates wolves from goats only on history)
my SB is already rationed for volatility...meditate on the picture, thanks ))
I'll let you know in private. so as not to incite trolls here)
I'll buy the grail.
Willing to give between $9 and $10, excluding transfer fees.
I'll buy the grail.
I'm willing to pay between $9 and $10, excluding transfer fees.
It's cheaper to throw away.
;)
I'll buy the grail.
The Grail... The great symbol of the suffering people of the world, including pygmies and Papuans.
I'm too lazy to write anything on this forum, but if anyone needs to talk to me, just shout out with a tear in your voice: "The Grail!" and I'm there...
The Grail... The great symbol of the suffering people of the world, including Pygmies and Papuans.
I'm already lazy to write anything on this forum, but if anyone needs to talk to me, all they have to do is shout out with a tear in their voice: "The Grail!" and I'm there...
The Grail... The great symbol of the suffering people of the world, including Pygmies and Papuans.
I'm already lazy to write anything on this forum, but if anyone needs to talk to me, all they have to do is cry out with a tear in their voice: "The Grail!" and I'm there...
this thread is very reminiscent of that movie: