The first IMF is [0,0,0....] all zero when set "parallel=False" using CEEMDAN or EEMD

[muqingCai]

I used the CEEDMAN like this ceemdan = CEEMDAN(parallel=False) ceemdan.trials = 10 max_IMF = 5 eIMFs = ceemdan.ceemdan(ts, max_imf=5) But get the result as
[[ 0. 0. 0. ... 0. 0., 0. ], [ 5.0205319 5.05441522 -4.62335232 ... 0.39283641 -6.48358469, -1.91464419], [ 3.3244119 3.34833722 -3.15371022 ... 0.73033395 7.22461791, 11.2976024 ], [ 1.61699314 1.6286....]

and when I set 'parallel=False', the result is true, the results are as follows: [[ 4.9391269 4.92000825 -4.62921163 ... 0.48549634 -6.37283759, -1.79111796], [ 3.60359462 3.61213567 -3.38193192 ... 1.11980536 6.91068809, 10.10722885], [ 1.32640539 1.34183321 -0.45469606 ... 14.26556025 17.83044674, 20.64068

I have used the latest code, the problem of "too many files" would be happened when used parallel=True, how can I get the true results when I use 'parallel=False'

[muqingCai]

@laszukdawid Can you solve it for me, thank you very much～～～

[laszukdawid]

Hi, unfortunately I'm a bit busy now. Most likely I won't look into this until the new year.

On Sun, Dec 22, 2019, at 06:45, muqingCai wrote:

@laszukdawid https://github.com/laszukdawid Can you solve it for me, thank you very much～～～

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/laszukdawid/PyEMD/issues/65?email_source=notifications&email_token=ACXNLK3T2J2DVQUKHPNQDRDQZ35G7A5CNFSM4J6KINF2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHPJFUI#issuecomment-568234705, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACXNLK425GKJFRVLMV6BQK3QZ35G7ANCNFSM4J6KINFQ.

[muqingCai]

Hi, unfortunately I'm a bit busy now. Most likely I won't look into this until the new year. … Do you think it's a bug or something wrong, my work has been blocked because of that. It's so sorry to bother you~~~

[laszukdawid]

I just read your first comment again and I'm not sure what's the configuration difference between both runs.

If you're getting an IMF consistently with all zeros, then that's a bug. However, it's more likely a bug in returning IMFs rather than with the algorithm. Please notice that your second IMF from the set with zero-IMF is quite similar (up to noise) to the first IMF in the set without zero-IMF. This suggests that you should be fine with simply ignoring the zero-IMF (and reordering the rest).

It'd be interesting to see if it's possible to reproduce the results. Would you mind sending me the data to pyemd@dawid.lasz.uk?

[muqingCai]

Thx so much, it’s very nice of you. I have emailed you and the data is as shown in the attachment. I think that ingoring the first IMF could get the right results but what caused the first imf all zero???

[laszukdawid]

3.5 years later. Unlikely this is relevant now. Sincere apologies. I'm only maintaining this package in free time and don't use it at all.

Seems that related issue had a similar problem and it's mentioned there that it's been resolved. It might be resolved here as well.