Numerical series density - page 14

New comment
 

Or is it like this: ?

10
13
15
21
40
42
46
51
56
65
71
78
81
 
Oh, no, it's the rows. I'll tell you from the table.

Can you tell me why you have 31 without a pair in the table?
 
Vyacheslav Kornev:
Damn, it's the wrong rows, I'll tell you from the table.
Will you tell me why you have 31 without -pair in the table?

How come 31 has a pair?


21 31 40 - on both sides the delta is greater than 6.
 
We have the following areas
10,15,13,21
40,42,46,51,56,
65,71
78,81
 
That's it, it's done. Densities of up to 6 and under are found.

If denser clusters are needed. Reduce the delta
 
Everything is variable. You may not count areas with less than 2 numbers as clusters at all.
 
Vyacheslav Kornev:
We have the areas...
10,15,13,21
40,42,46,51,56,
65,71
78,81

It's not clear yet how you grouped them - by eye - or by the same principle as I did in my algorithm.

OK, this was all visible from the table as wellhttps://www.mql5.com/ru/forum/163871/page11#comment_4131070

However, my algorithm eliminated 65 and 71, and in this case that's good, but maybe there are numbers on which my algorithm will be worse than yours - how to check that...

And, here we have areas - which one is denser - according to your view?

 
Or is it wrong? The task was to find clusters, wasn't it? Our numbers are just ordinal numbers. It's only the delta that counts.
 
The tightest is 40 and 42
 
Look, you can also cut 65 and 71

Take delta 5

Everything else will be there.
1...789101112131415161718192021...26
New comment