There is difficulty in defining in what a fractal is because there are many fractal dimensions (boxcounting, correlation etc.).
Quite interesting.
Is something still a fractal when there exists a fractal dimension which is equal to topological dimension and a fractal dimension which is not equal to topological dimension?
🧐 Why does this sound contradictory. Or are you talking about an object with 2 topological dimensions and above?
Probably in the first dimension of, say a 2D object, there's self-similarity but it's topological dimension is not the same as the fractal dimension and in the second dimension there's also self-similarity but the topological dimension is the same as the fractal dimension. Or is there something I'm missing ?
If it's for the 2D object, then i don't know. Plus i would love to see a visual representation of this kind of pseudo-fractal - if that's the right word to use.
So we can just take the real line and select the rationals over that. Then the topological dimension is 0, hausdorf dimension is also 0, but the box-counting dimension is 1.
For adding coordinates you can of course extend this argument to n-dimensions. :3
So you are saying that supposing we had a set S with points/elements (rational numbers) such as {1.1, 1.2, 1.3, ...} - gotten from a real number line, we could pick any point, say 1.1 and scale/magnify it down by 0.1, and correspondingly we have points which is also a set, given below as
X = { 0.11, 0.22, 0.33,...., 1.1}
X has cardinality of 10
Which means that every point in S when scaled/magnified by a non-zero factor gives a new set of points ( e.g X for 1.1), and every point in this new set also gives it's own corresponding new set when also magnified, and so those the process continues to ad infinitum. The self-similarity here is that points are contained in a point when magnified. However, a single point has a topological dimension of zero which is equivalent to the point's fractal dimension but the fractal dimension (box-counting dimension) of any magnified point, say 1.1 in S above is given by
D = log(N)/log(1/e)
D = fractal/box-counting dimension
N = number of points in the magnified point (1.1)
e = scale/magnification factor.
In our scenario as seen above
N = 10 (cardinality of set x)
e = 0.1
Therefore D = 1
This seems to me like a fractal.
It works more general for all the rationals. I think it is true that any subset of the rationals gives a box counting dimension \leq 1
The main point is that the Hausdorf dimension is zero. Because it is countable. Thefore, the Hausdorf dimension says it is not a fractal whereas the Boxcounting dimensions says it is a fractal.
Take it easy comrade, I'm aware
It (subset of rationals) actually does.
I think I get where you are coming from.
It's nice learning from you anyways, thanks.
I am happy to share :3
Ah yes, this is obviously true :P
An interesting exercise would be to construct a non-finite subset of the rationals with box counting dimension equal to zero :3
Ok, good luck with that.