Siniara

Thank you for the reply! It's a very interesting point you make about the constant subsequences that I hadn't noticed mentioned in the related works for the method. Do you...

I really appreciate your informative answer! I'm a Master's student doing a project with the SAX transformation, so I'm just learning about this method. Indeed, I'd missed the bit about...

Good catch! The time series I'm using is relatively long to share, when I try to make it smaller/downsample I don't get the same patterns anymore (unsurprisingly), so it's hard...

Your help is very much appreciated! For my problem the interpretability is key, hence SAX was a method I considered as then I could look at the patterns that are...

Thanks for the fix! So now indeed the SAX-VSM doesn't throw that error anymore. However, when I tried to use the approach you suggested: > Use pyts.multivariate.transformation.MultivariateTransformer with pyts.bag_of_words.BagOfWords, which...

Yes, I installed it as you outline, my current version is `pyts 0.12dev0`. The code snippet you provide works, but I just retried on my data and still the message...

My X.shape was `(2396, 7, 51)` -> number of multivariate samples/time series, 7 features and 51 measurements per feature. Which seems to correspond. I also tried a longer time series...

I just tried to apply `BagOfWords` separately on each feature as you suggested and also got the error. This is the original error message from using the Multivariate Transformer. Full...

For 3 of my features I get the constant sample error and for 2 others I get this error: ``` 176: UserWarning: Some quantiles are equal. The number of bins...

I just copy-pasted your code and provided my data as X, since I don't see that parameter explicitly stated it should take the default value. And in my previous experiments,...