AR1 Warning: Cannot place an upperbound on the unbiased AR(1). Series is too short or trend is to large.

[shenghwa]

When I want to replace the nan values of the dataset with mean value, sometimes it will raise the warning. I can't tell the difference between the problematic data and the normal data. If I use zero or mean/2, this warning will not raise. Does anyone meet this issue?

[claytonjschneider]

I have this issue as well

[claytonjschneider]

I also see a slight discrepancy between the github master repo and the latest pip install (from today), where the 'if sig:' line in wct occurs after the autoregressive function. So even though I'm giving wct sig=False, it triggers, and fails

[claytonjschneider]

I know this is absolutely horrible practice, but I'm going to move the if sig line up above the ar calculations in my pip's version of 'wavelet.py' and see if that fixes it for me, since I'm not concerned with the significance values at the moment (though I would like to use them at some point).

[claytonjschneider]

Seemed to do the trick.

[ranzan998]

@claytonjschneider can you please tell if there is any better solution to this?
In case you made a workaround. If yes, please let me know.

[claytonjschneider]

@claytonjschneider can you please tell if there is any better solution to this? In case you made a workaround. If yes, please let me know.

Sorry - haven’t been involved in cwt since my original comment. I remember that workaround fixing it, but not sure whether updates in the last 2 years have made it obsolete.

[ranzan998]

Sorry - haven’t been involved in cwt since my original comment. I remember that workaround fixing it, but not sure whether updates in the last 2 years have made it obsolete.

I see.

For cwt, I could get around if the data is long enough. But for coherence transform wct function, it stops working even with the data records whose length was sufficient for cwt.
I read somewhere it's not maintained anymore.