Discussion of article "Statistical Probability Distributions in MQL5"

New comment
 

New article Statistical Probability Distributions in MQL5 is published:

The article addresses probability distributions (normal, log-normal, binomial, logistic, exponential, Cauchy distribution, Student's t-distribution, Laplace distribution, Poisson distribution, Hyperbolic Secant distribution, Beta and Gamma distribution) of random variables used in Applied Statistics. It also features classes for handling these distributions.

Author: Dennis Kirichenko

 

Very interesting, thank you, good work. If it is not difficult to add the calculation of the distribution of a function given tabularly, so that there would be something to compare it with.

And in addition, a method for determining the greatest similarity with theoretical distributions (this can be done through the correlation coefficient).

 
Urain:

Very interesting, thank you, good job.

Thank you for your opinion.

If it is not difficult to add the calculation of the distribution of the function given tabularly, so that there would be something to compare with...

Please specify. Better on an example :-)))

And in addition to it, a method of determining the greatest similarity with theoretical distributions (this can be done through correlation coefficient).
What do you mean? To what extent does the empirical distribution differ from the theoretical one?
 
denkir:

Thank you for your opinion.

1) Clarify please. Better with an example :-)))

2) What do you mean? To what extent does the empirical distribution differ from the theoretical one?

1) A function given tabularly means that there is a data set (e.g. an array) where each x corresponds to y, but the dependence formula is not known.

Such a function is in fact quotes. And that's what I'm talking about: calculating the probability distribution of such data.

2) Yes. Which of the theoretical distributions is more similar to the empirical one. Or just the correlation coefficient between empirical and theoretical.

 
Urain:

1) A function defined tabularly means that there is a data set (e.g. an array) where each x corresponds to y, but the dependency formula is not known.

Such a function is, in fact, quotes. And that's what I'm talking about: calculating the probability distribution of such data.

Either I misunderstand something or... usually in the tabular form, already known theoretical distributions are given. Personally, I don't like tables very much. I can see better on a graph, so to speak... and I can see the shape of the distribution... In the video shown in the article, you can see how the values change when moving the cursor. And this is only one way of representing the distribution law... you need a lot of tables to cover everything... and a graph can.....

2) Yes. Which of the theoretical distributions is more like the empirical distribution. Or just the correlation coefficient of the empirical and theoretical.

In the conclusion of the article, I wrote like this:

I, for my part, am going to develop this topic and demonstrate with practical examples how statistical probability distributions can be used in analysing probabilistic models.

More details a little later.

 
denkir:

Either I'm misunderstanding something or.... usually in tabular form, already known theoretical distributions are specified. Personally, I don't like tables very much. I can see better on a graph, so to speak... and I can see the shape of the distribution... In the video shown in the article, you can see how the values change when moving the cursor. And this is only one way of representing the distribution law... There's a lot of tables to have to cover everything... and a graph can....

In the conclusion of the article, I wrote this:

I, for my part, am going to develop this topic and demonstrate with practical examples how statistical probability distributions can be used when analysing probabilistic models.

More details a little later.

No no, you don't need to draw analytical functions as a table, I meant to create a method (programme function) to calculate the probability distribution of quotes. Quotes is a function defined tabularly, without knowing the formula by which x to y conversion takes place.

OK, let's wait for the continuation.

 
Urain:

No no, it is not necessary to draw analytical (defined as a formula) functions as a table, I meant to create a method (programme function) to calculate the probability distribution of quotes. Quotes is a function defined tabularly, without knowing the formula by which x to y conversion takes place.

OK, let's wait for the continuation.

Ah, well, this is called fitting to the theoretical distribution, if I caught your thought accurately... I'll talk about it in detail later... in practice with a few examples... especially since there was a heated polemic about distributions when discussing my paper:-))
 

One of the greatest articles on MQL5.com community!

Thank you very much, Dennis!

New comment