Machine learning in trading: theory, models, practice and algo-trading - page 1729
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
take a pair of components, for example 3, 4
on the x-axis - component 3
on the y-axis - component 4
If it's round - the loop
That's all)) Man, I thought there was some kind of cross correlation or something.
That's all)) Damn, I thought some kind of cross correlation or something
Correlation counts too.
That's all)) Damn, I thought some kind of cross correlation counts or something
Karoche the meaning of my message was this. This non redrawing crawler has a number of settings, where you specify parameters from the green window 10 and 11yo and set their groupstar and groupdepth, then you select the range of construction (important) and the number of bars of training I do not understand and never changed it, I think it has something to do with iterations. At the end it allows to build a cyclic line of absolutely any curvature and cycles. This thing is very flexible. You know, a caterpillar that doesn't redraw. So the point of this strategy is that when you build this straight line according to the parameters of the green window, which you also select manually. The dark stick on the points expands and moves along the points, and at that moment the oscilloscope is dancing like hell. In this case, the oscilloscope will move to the points and move the points manually at that moment. But the most interesting thing is that this addon was the only addon that had its own window inside Neuroshel. Not a dialog form of induke or strategy, and it opened the internal window of the program along with the main form, it's not an induke or strategy is kind of a tool for work. I had a great idea how to use it in order to help me cope with my opponent's situation. Well krochet dance with such a low probability of finding this circle, I did not, but the circle that I got is not chetamoyu dull placezna on the green window. Look, even the creators of this tool couldn't find a good example, because there are a lot of variants of this curve at the current chart end and it will describe it like a caterpillar, but in the nearest future it will behave differently. So in the green window there are two frequencies, which alternately predict each other.
The fundamental point of this strategy is that it is guaranteed to give 4-5 signals immediately after the construction is guaranteed to take its own and fuck off. Yes, even two will be enough. This is the fundamental meaning of all strategies. If a strategy does not give you a guaranteed 5 signals immediately after optimization, then you have no strategy. Well, even if it gives 3 out of 5 in the plus it will already be enough.
So in this strategy is the whole point that when you have built and got a circle. Here and now you are guaranteed to have three moves for which you will, in any case, earn money, make them and go further. It's such a tool. Just when I got a circle of the right shape and it was a spiral or rather if you draw two spirals in each other I okue, it really did not work long within 5 bends where the first three were guaranteed, I do not remember exactly. But unfortunately it was just not possible to do it by hand. Although I was a concrete nerd on the Neuroshell, more than one mouse was replaced, :-) but on such a test was not ready. So Maxim, ask me about the NS, I'll explain to you specifically for him. :-) Not ..... programs fire, in fact.
I have explained everything to you.
The correlation also counts
Well, I see, it's part of the PCA.
I was just wondering how to build these two-dimensional diagrams, just try to feed them AMO recognition, because they(two-dimensional diagrams) are considered dimensionless and express the dependence of one component on another, so you can compare each with each and we describe the entire series, and dimensionless those withoutabsolute values of the series. I am interested in your thoughts and criticism
Did you write the program yourself?
I think that all this will not work, the price changes its characteristics very much, I "work" on something that will transform the data correctly even before all the manipulation, which will be looking for some statistical patterns, I think that this is goal number 1, if it will be done then it does not matter which MO to apply or type of optimization
look at the root, not the square of the OLS
this is the transformation of the data, in which some component is stored, for example, the mean of returnees
it's analogous to clustering, if you will.
you started out doing good stuff, but then suddenly got dumbed down and turned sideways, imho...
Did you write the program yourself?
there's a good tutorial
There was also something that said that you can see which frequency predicts which frequency and since they both belong to the same series. Well, in general the curvature turns out to be preemptive on literally 2-3 kinks. What is not a strategy and most importantly no neural networks, stable and guaranteed here and now. In general it would be great to check this in general, maybe some additional manipulations could be invented to ensure the correct settings that lead to the curvature being ahead of the kotir for some time in the future. Questions?
Tried 10 years ago caterpillar, fish are not there in the classic form
SOFTWAREyou look at the root, not the one squared from the ISC
this is the transformation of the data in which some component, such as the mean of the returnees, is retained
it's analogous to clustering, if you will.
You started out doing good stuff, but then you blunted and turned sideways, imhoI don't get it. Explain the root and where I went wrong.
There's a good tutorial.
Thanks for the tutorial...
Did you write the program yourself?