Machine learning in trading: theory, models, practice and algo-trading - page 2086
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
Well... Zircon is fun... a lot of fuss, of course, but here's what happened last time
It's worth poking around.
But if you add a spread :D
what is zircon?
What is zircon?
https://pyts.readthedocs.io/en/latest/auto_examples/transformation/plot_rocket.html
At first I thought this thing could fight overtraining
But then it became clear that it was just a trait transformer... a good transformer. But I want less overtraining.
Look, the MA(moving average) is a simple filter, for the MA(4) its IX (0.25, 0.25, 0.25, 0.25), for the MA(5) (0.2, 0.2, 0.2, 0.2).
For the MA(5), we take the last 5 prices, multiply each one by 0.2, add them up and this will be the value of MA(5) on the last bar.
You don't need it now, just look at the first part of the article and understand the filter types.
Thanks, you really explained it, I can't understand formulas(
I read about filter types long ago, I understand in principle.
I'll show you an interesting thing, if I can reproduce it.
I searched my archives, found it, remembered it, figured it out, but it's still an interesting thing
The target is an approximated series, I approximated "ssa", but Fourier is also possible, the main thing is to smooth it out nicely
Like this
Then in the sliding window
do the Fourier transform, remove the harmonic with zero frequency, find the harmonic with the largest amplitude (as you said)
and give it to the MSUA to study
after the training we get an incomprehensible but very clear signal, you can clearly see that the harmonic
And here at the same data Forest did not find anything
I think how to distinguish such clean and useful signals from the spectrum ?? Very interesting thing this spectral analysis ... and filters on it...
By the way a couple of pages ago they were saying that Forest can't interpolate data inside itself, but the grid can, here's a vivid example
you canhttps://www.mql5.com/ru/forum/244716#comment_7451342
the code worked before without problemshttps://www.mql5.com/ru/forum/123222/page4#comment_3229236
It depends on what we are looking for, if we confirm that during the news there is a surge of valotility (the code at the link looks for it), then yes it does, because the news release is a certain ritual during which there is a ban on the placement of orders and... the legend? the big strong players remove their orders during this action, which leads to a spike in valutility, the source of legend - the Internet resources of various forex brokers
if to confirm that the news unfolds the market - I doubt it is possible, then ZigZag with a large step to help, trends often unfold at night, when no important news for everyone was released
Recently posted some similar code here - with volatility calculation depending on time of day.
That's the thing - simple volatility change doesn't give an opportunity to earn money (except with options). Some predictable movements are needed.
I recently posted some similar code here - with volatility calculation depending on time of day.
That's the thing, just changing volatility doesn't give you an opportunity to make money (except with options). You need some predictable movements.
From the point of view of physics, we need to look for impulse action. Volatility does not depend on a single impulse. It will not change. On the other hand, you can look at what happens to the price, and for how long. I don't believe that nothing happens at news time.)
Date, time is, the rate and range of change in the price of the pair in the time frame of half an hour before and 3 hours after the news. The news are given for the country. The first archive breakdown by country and type of news. Select the country and view the effect of each news item on price. I can't think of anything else.
In addition there is a forecast of the importance of the news M L. It will be possible to compare. I do not know what the numbers after the letters of significance mean.
It's New York time) and I have to take into account daylight saving time.) Or skip the day of transition.
Transitions are there - among the news marked with N
Since it's parsing FF calendar, first three numbers - value, forecast, previous revision; the fourth - I don't know, may be the previous revision before it; last integer - from chart address.
Themacd is a bandpass filter + low-pass filter from it. We get cutoff frequencies from the spectrum - 2 parameters, we take the signal line arbitrarily, it adds smoothing and delay
And if you think about it?
A bandpass filter is a filter that passes a specified range of frequencies (band)
to do this, you must first at least perform the Fourier transform
the physical meaning of MACD is a delta between two price averagers, each oriented to a different time interval, and they are compared now
From the point of view of physics, we should look for an impulse effect. Volatility does not depend on a single impulse. It will not change. On the other hand, we can look at what happens to the price, and for how long. I don't really believe that nothing happens during news releases.)
Obviously, there are changes, but the opportunity to make money on them must be established. This is the standard way of matstat reasoning (hypothesis test) - we make the null hypothesis and its alternative, and then test them with the data.
Karoch )) do not need anything, no frequency or phase, just feed one amplitude of one harmonic with the largest amplitude, just one sign
if you add something, you get a lot of noise
and you get a very clear signal, it is another question what this signal has in it))
The changes are obviously there, but the possibility of earning on them must be established. This is the standard way of mathematical reasoning (hypothesis testing) - we put forward the null hypothesis and its alternatives, and then test them with the data.
I see the task as more difficult. Taking into account only the news is not meaningful, the experience of traders proves it) It is necessary to watch the price change, but I would look for significant changes in price from the news to find what else affected the price. Watch what news was around. And maybe mix it in with stock indexes or whatever else is in the number.
In addition, describing the news only in terms of significance is flawed. There is an expected value of the news, there is a real value, and there is a direction of influence on the price.