lrydin / kramersmoyal Goto Github PK

Dear authors,

I'm sorry to pollute the github project, but can you comment a little bit on the bandwidth parameter?
It's the only free parameter and it seems to have an important influence on the results, but it is not discussed in the docs nor on the paper.

Thank you for this great work,
best

Hi I am interested in your computing multidimensional DFA such as that found here: https://journals.aps.org/pre/pdf/10.1103/PhysRevE.74.061104

I noticed in your examples that it specifically mentions using 1d. Can your MFDFA support 2d?

I ran into the following error while executing one of the tests in test directory. Maybe I'm missing something and the authors can either fix or add instruction to specify usage. Other tests ran just fine.

run test/binning_test.py

TypeError Traceback (most recent call last)
~/Documents/review-JOSS/KramersMoyal/test/binning_test.py in
15
16 hist1 = [np.histogramdd(timeseries, bins=bins,
---> 17 weights=w, density=True)[0] for w in weights.T]
18
19 hist2 = histogramdd(timeseries, bins=bins,

~/Documents/review-JOSS/KramersMoyal/test/binning_test.py in (.0)
15
16 hist1 = [np.histogramdd(timeseries, bins=bins,
---> 17 weights=w, density=True)[0] for w in weights.T]
18
19 hist2 = histogramdd(timeseries, bins=bins,

TypeError: histogramdd() got an unexpected keyword argument 'density'

One of the JOSS requirements is to have community guidelines and code of conduct for interacting with the project. Authors can link pointers from the README to options under insights about:

Contribute to the software
Seek support for usage (authors have added this in README and can link it to the guidelines)
Report issues or PR (if the authors think users need to report in a specific style)

There are helpful templates created by GitHub to do these in Insights from the bar just under the repository name.

(Part of openjournals/joss-reviews#1693)

First, thank you so much for writing this package! Being able to quickly call a function to compute KM coefficients has been a considerable help in my current research.

I have a clarification to suggest in the main README.md regarding the normalization. Currently it is stated that one should multiply kmc by delta_t to obtain the actual Kramers-Moyal coefficients, but if I understand correctly one should in fact divide by delta_t.

Overall the paper reads well but there are a number of typos to be fixed - see the attached PDF. (There is a little to much enthusiasm for commas though.)

10.21105.joss.01693-annotated.pdf

Hello, I came across your package in my effort to replicate the methodology from paper:
https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1000061
where in the Data Analysis section they use the Kramers-Moyal coefficients to get the drift coefficients of the vectors and subsequently the angle of neighbor vectors in the phase space in order to make conclusions about the dynamics of the flow in the state space.
I don't know if you would be interested in helping me a bit with the code. I actually want to do it for a 3d space, but even in 2d i don't quite understand how to get the angles from the output of the km function since i get this bins x bins x bins vector.
I hope I was clear. Thanks a lot in advance and apologies if my issue fall out of the scope of the issues section :)

I'm just going through your submission to JOSS and following through your readme I can't reproduce your results exactly.

1. The code errors since kernel.epanechnikov is an unknown name (it should be kernels.epanechnikov).
2. The code errors saying that the first power should be zero.

Fixing those two mistakes I then get the results that I plot using

plot(edges[0], kmc[1])  # the first coefficient

There is some distortion at the extremes (I presume to be expected due to low numbers of samples?) and it shows the linear trend demonstrated in the readme figures but the values are a factor of 1000 different.

A couple of minor points that don't affect the calculation but are more about style:

1. y = np.zeros([time.size]) could simply be y = np.zeros(time.size), and
2. size=[time.size,1] could simply be size=time.size.