search

lucasfr / chhabra-jensen

Multifractal analysis in Matlab and Python
MIT License
7 stars 0 forks source link

Regarding the scale value #3

Open thediffigral opened 2 years ago

thediffigral commented 2 years ago

First of all, thank you so much for the code. It really helps to understand the topic. But when I tried to run the code for my time series, it shows some errors. Mainly because my input time series file contains 1000 values. How to make the code run for time series of any length?

lucasfr commented 2 years ago

Hi @thediffigral Thank you for your comment. I am happy the code is helping... The way this algorithm has been implemented will not allow series of any size as it's based on the dyadic scale for the window sizes. I would recommend you (if possible) getting a bigger time series and either trimming 2^n sized epochs or using sliding windows (also with 2^n sizes). Using epochs of 512 values might help in you case, though you will be losing a scale and perhaps making the power law fitting not ideal.

thediffigral commented 2 years ago

Thank you so much for responding to my query. I am taking note of your solution. Just if you have some more time, I have the last query. Looking at the fashion of using window sizes (2^n), is it the logarithmic binning? Are you using logarithmic binning to evaluate the probability?

Thank you in advance :) I hope I am not disturbing you.

lucasfr commented 2 years ago

Hi @thediffigral Sorry for the late response. I ended up missing your comment. If I understood your question correctly, yes! We fit a power law for the measures distribution.