Closed LRydin closed 3 years ago
First draft at 92fa429b3941f0b9be26f264326a7eab2b8d341f. I am pondering designing an adaptive moving window, as the problem scales with the lag
(or even the sum(lag)
), making it rather slow if there are large lags to study.
On the other hand, I have given priority to loops to avoid allocating more space in the memory. The way it is designed now should not require any extra memory compared to the conventional MFDFA
.
Adding moving window to dev
state in the pull request #14.
Closing #4, in issue #17 a moving window in implemented. Now out in v0.4.
Implement a moving window to analyse shorter timeseries or improve results at large lag times.
At this stage I don't recall where I first read this, so I have to find the original publication and add it later.