Econometrics: let's discuss the CU balance sheet. - page 15

New comment
 
Avals:

just means not enough deals for Koshi to demonstrate himself))
again -- that's not the point !!!
 
avtomat:
But then again -- that's not the point !!!

show the distribution of the rest of this row please
 
avtomat:
But then again -- that's not the point !!!
Yes? Tell me where your old tractor is then?
 
TheXpert:
Yes? Tell us where your old tractor is then?
The question is improperly posed. And inappropriately so. The tractor is developing - it's a long process. Now a fresh version of the tractor is in operation (I don't know which modification, I don't keep count of such modifications due to its uselessness). The last experiment was terminated not through tractor's fault, but due to force majeure. However, I am sure that many people would have been pleased with the results of the tractor ;)))))))
 
I'm telling you, he's a charlatan.
 
Avals:

show the distribution of residuals of this series please

It is also interesting to see the distribution of e.g. 1000 generations of such series. I.e. to montecarry
 
TheXpert:
I'm telling you, charlatan.

Yes, quack ;))) because I reject normality as a criterion ;)))) and it is a sacred cow ;)))

 
Avals:

it is also interesting to see the distribution of e.g. 1000 generations of such a series. I.e. to montecarry.
specify the question, because I have missed what exactly you want to see
 

A vector of independent random variables that has a Cauchy distribution:



 
avtomat:
clarify the question, because I missed what exactly you want to see

If you generate a lot of null motes and combine all the trajectories on one graph, you get a bell)) If a sub with a drift, then a bell with an upward slope. This will be the pattern when the mo is constant and the residuals are normally distributed. This picture has been posted on the forum many times, it has also been reworked into a Christmas tree)). If you also generate with Cauchy distribution of residuals, then clearly there will be an inevitable margin call in almost all implementations depending on row length. Of course, it is possible to invent a huge trend component and Cauchy with a minimum coefficient of scale (or whatever it is set there when it has no variance) - but this is fantastic. And still for a long series there will be MC
1...8910111213141516171819202122...30
New comment