Tony Bagnall

Hi @fkiraly I'm housekeeping, I want to reduce the number of open PRs by deleting stale ones. This PR is conflicted and has been inactive for a long time. Could...

@fkiraly is there still an "unofficial" distinction in policy for the deprecation policy for stable/maturing/experimental modules as defined here? https://github.com/alan-turing-institute/sktime/blob/main/README.md in the past we have not required a full cycle...

it definitely was the case that experimental did not require full deprecation, I remember distinctly with the clustering module. Thank you for clarifying that this is no longer the case...

I have not seen this sporadic fail, and the linked logs have expired, Can we close this @fkiraly?

not a full list, I'll sort that after lunch, but just posting quick summary metric str: default to accuracy, others include nll, f1, auc resamples: int - the number of...

hi chris, I will rearrange the file structure later, but for development you can try loading this file https://timeseriesclassification.com/results/ResultsByClassifier/TSF_TESTFOLDS.csv

and the other thing is where to put this. Just chuck it in contrib in the first instance, I may ask a student to do a workbook based on it

we have some prototype code to do this, but on reflection I don't think we want this in sktime proper, since it will introduce a dependency on an external website...

we have a dependency on tsc.com for loading data from the website, but somewhere you said you were working to disconnect sktime from any UEA infrastructure, which includes tsc.,com, so...

I would have thought people working on it might comment on your issue or might be assigned. I don't think the original contributors wanted to do it, hence why I...