Thanks to @marielledado for suggesting this.
Thanks @mloning !
I'll start working on 02_classification.
Basically I plan to "preserve" the code from the tutorial notebooks and only take relevant code snippets & explanatory text from the code breakfast notebooks.
03_classification doesn't seem to use the gunpoint dataset. Now working on feature_extraction_with_tsfresh.
Oh yes, my bad, that's on multivariate data only!
@mloning Would it be okay to also replace references to GunPoint in the loading_data notebook? This would involve adding new arff and tsv files, which I found for the ArrowHead dataset.
A related question would be, would there be any use for the GunPoint dataset after making all these replacements? If it's essentially obsolete then may be it should removed from the library altogether?
@marielledado Yes, feel free to replace it there as well.
The only notebook where we use the gun point data set is to explain shapelets based on the published paper and diagram. My impression is that removing it there takes away of the intuition what the algorithm does. Any ideas?
@mloning I did some digging around and found a paper that used shapelets on a reptile skull dataset. The figures are pretty illustrative, but I couldn't find a dataset for it unfortunately. I suppose it would be good for the sample dataset and the illustrations to be the same in the tutorial.
I suppose we could leave the gunpoint in the shapelet notebook for now, then maybe I can email the authors to ask if they can provide the reptile skull dataset?
@marielledado Nice work! Perhaps you can find it in timeseriesclassification.com? One of the authors is associated with the people who maintain the website and data set repo.
Update: I had a quick look but couldn't find it, seems to be too small to be included.
Update 2: The leaf example here may be a good candidate: https://www.cs.ucr.edu/~eamonn/shaplet.pdf There's a leaf data set as well (even though a different one): http://www.timeseriesclassification.com/description.php?Dataset=OSULeaf
@mloning Nice, I'll give the leaf data set a go.
Closed by #295