Numerical series density - page 23
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
now back to the very beginning of the thread again :-)
"what is the density of a point" ?
Don't be so exaggerated :)
I can see that there is a definite solution, but this method emphasised a particular group that may not satisfy the parameters of the number cluster area being sought...
About density - I see two possibilities:
1. (NumberStart of Row-NumberStop of Row)/Number of Numbers.
2. SumTruth/NumberTruth
The first option emphasises auniform distribution, the second option emphasises a relative distribution (let that be the word).
The task was to find dense clusters of points. To do this we took the density and actually derived it, i.e. we got the derivative. Based on the derivative we can say "here is the maximum", "here is the minimum", the density increases here, and slowly decreases here.
But we cannot compare absolute values - to do so we need to calculate the original function (in this case we simply take and count the number of points in some vicinity of extrema).
Yes, the approach is interesting - thank you.
Perhaps it will also prove to be excellent - but it's a whole story for me to program it all - without testing on large numerical series it's too early to make final conclusions about whether this option suits me or not.
Something else you are discussing.
About 1 - you understand that there can be almost as many different deltas as figures themselves - in this case the solution is not productive, because you cannot know beforehand by what criterion (how many deltas to take) to group the figures. Can't you understand that?
About 2 - yes - such variant of the solution is understandable - as an opinion on the problem.
About 1 - you realise that there can be almost as many different deltas as there are numbers - in which case the solution is not productive, because you can't know in advance which criterion (how many deltas to take) to group the numbers by. Can't you understand it?
About 2 - yes - such variant of the solution is understandable - as an opinion on the problem.
What the hell. We may have deltas 1,2,3,4,5,6,7 at least. Accordingly you will find clusters in order of density.
That's what I suggested long ago - find clusters in order of density and find the density of each one separately, then compare them.
But, I saw that as the density grows, left-hand figures start to fall in - which noise up the clouds - so I abandoned this idea.
But, I don't have a tool for conducting a large number of experiments - your method should be programmed to be able to compare - I'm not ready to do that now - I have no experience of working with multidimensional arrays.