Machine learning in trading: theory, models, practice and algo-trading - page 2849
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
Roughly speaking, you're not too lazy to fell trees, but you are too lazy to sharpen your axe.
Risk function, the simplest variant on R
The sections of the curve close to the horizontal line correspond to dips in the histogram, and here these sections can be determined more precisely, since there is no connection to the partitioning (as in histograms). I use, for example, when studying the distribution of heights of zigzag knees.
Sorry for possible misunderstanding of the question.
Or is it from another topic?Can the Huber function be considered as a risk function?
It seems to be calculated as you have shown on R.
Only as I understood it defines a 10% percentile for emissions.
Is it possible to apply the Huber loss function as a risk function?
Decreases from nx to 1. For example 5:1 = (5,4,3,2,1) and 1:5 = (1,2,3,4,5)
As usual in matstat - an empirical analogue is constructed from the sample. Like mean instead of expectation, frequency instead of probability, or ECDF instead of CDF.
Okay, so I've plotted it, what do I do with it?
And how can I use a histogram if x is the number of elements in the sample?
And how can we make a histogram if x is the number of elements in the sample?
X should be a sample (of column heights, in your case) sorted in ascending order. And the function should be increasing from zero to log(nx). If, for example, nx=5, then y=( log(5/5), log( 5/4), log( 5/3) , log( 5/2) , log( 5/1)) .
Or is it from a different opera?
Absolutely different. You are talking about one of the variants of the loss function, while we are talking about the cumulative hazard function.
And reinforcement training?
What are you going to reinforce with?
Anyway, you won't have any intelligence, but only a model of the autonomic nervous system, and that ...
Maybe you'll be able to practise some reflexes ...
And where is it, intelligence? Where are the levels of abstraction? Where's the schizophrenia?
Where is all this in your artificial intelligence?
For X there should be a sample(of column heights, in your case) sorted in ascending order. And the function should be increasing from zero to log(nx). If, for example, nx=5, then y=( log(5/5), log( 5/4), log( 5/3) , log( 5/2) , log( 5/1)).
Important clarification!
Is it like this then?
And how to transform the histogram then?
Important clarification!
Is that so?
And how to convert the histogram then?
Well, you can already see the horizontal sections. It is also disconcerting that the maximum sample you have here is 400, while before it was about 60. Perhaps you should take log(X) instead of X, having previously thrown out zero values from the sample - this will allow you to see the area of small values of X in more detail.
Anyway, I don't know what your task is in general. The method only answers one specific question you have - how to separate the highest "fence" from the lowest "trees". The beginning of a horizontal section (or close to horizontal compared to the average slope of the rest of the curve) is the highest fence, and the end of such a section is the lowest tree. There are either no or very few points on this section itself, which allows us to neglect them.
Well, you can see the horizontal sections now. It is also disconcerting that the maximum sample you have here is 400, while before it was about 60. Perhaps you should take log(X) instead of X, having previously thrown out zero values from the sample - this will allow you to see the region of small values of X in more detail.
I have transformed X, but I still don't understand what you saw there - and how to automate the process of defining coordinates on it to select the desired range.
Can you name the specific coordinates where the "gentle graph" starts? And there was a horizontal graph, and then movement at a sharp angle - it doesn't count anymore - until the first gentle graph or what?
In any case, I don't know what your problem is in general. The method answers only one specific question of yours - how to separate the highest "fence" from the lowest "trees". The beginning of a horizontal segment (or close to horizontal compared to the average slope of the rest of the curve) is the highest fence, and the end of such a segment is the lowest tree. There are either no or very few points on this section itself, which allows us to neglect them.
The goal is to find predictors describing the nature of the sequences.
Converted the X as well, but I still don't understand what you saw there - and how to automate the process of defining coordinates on it to select the desired range.
Can you name the specific coordinates where the "gentle graph" starts? And there was a horizontal graph, and then movement at a sharp angle - does that not count anymore - until the first gentle graph or what?
The goal is to find predictors that describe the nature of the sequences.
In the first figure, the obvious horizontal plot is from about 2.4 to 3.
If it were, for example, a sample of the heights of the knees of a zigzag, this is an opportunity to enter on the breakdown of the first level and take profit on the second.
If it were, for example, a sample for the lifetime of an arbitrage opportunity, it is better to enter those that survived to the first level.
There is no strength, time or desire to think about how exactly you can use this curve. I have already told you many times that I am an opponent of the idea of "joint work" on the forum. I see the benefit only in a superficial discussion of individual theoretical issues.
I am an opponent of the idea of "working together" on the forum.