Multifractal estimation---maximum likelihood
by Emeritus Prof. A. J. Roberts
Home page
Linear Algebra for 21st C
Equation-Free Toolbox
Modelling emergence book
Financial math book
best LaTeX introduction
Holistic discretisation 1
Holistic discretisation 3
SDE slow manifold et al.
Construct invariant manifolds
Lego Fractals
Some software
Research publications
Amazon's Author Page
Other homes
Mathematical Sciences
Faculty ECMS
University of Adelaide
Contact me
Questions marked by * are required.
Name: *
Email: *
Message:
Due to a server upgrade some years ago, this web service is not working at the moment. Contact me by email.
Paste your data points into the box below, then click
Also enter the magic word "a z o n a l" into this box:
Enter your data into the box to the left
Your data must represent a 1D, 2D or 3D set of points.
Your data must be
one, two or three
columns of numbers;
and no more than
400
rows;
then click the submit button above to analyse.
Wait for a minute
Your data is analysed after submission:
taking up to perhaps a minute execution time;
this is compute intensive, so
have patience;
and you will then
see a data summary returned in a web page.
The theory behind the analysis
A. J. Roberts and A. Cronin,
Unbiased estimation of multi-fractal dimensions of finite data sets
,
Physica A
, vol.233, pp.867--878, 1996.
I construct many artificial multiplicative multifractals and remember those that generate fractals that look much like your data. The reported fractal properties are those of multifractals that best look like your data.
Be wary that the Hausdorff dimension is very sensitive to experimental sampling chance: this article explains,
Use the information dimension, not the Hausdorff
I thank Thuc Le for the code to handle data embedded in different dimensional space.
If you like this web page, please link to it so others can find it more easily.