I found an issue when working with the ts file format - if I was using unlabelled data and so defined @classlabel false in the header section of the file then it would throw an exception when running. The issue is that the return from load_from_tsfile_to_dataframe will return a tuple when @classlabel is true, but just the dataframe when @classlabel is false. Therefore, when unpacking the expected from load_from_tsfile_to_dataframe it would throw an exception, and fall back to your implementation, however this assumes that the last column is a label and you end up with a truncated data set.
Thanks for this code and the paper!
I found an issue when working with the ts file format - if I was using unlabelled data and so defined
@classlabel falsein the header section of the file then it would throw an exception when running. The issue is that the return fromload_from_tsfile_to_dataframewill return a tuple when@classlabelistrue, but just the dataframe when@classlabelisfalse. Therefore, when unpacking the expected fromload_from_tsfile_to_dataframeit would throw an exception, and fall back to your implementation, however this assumes that the last column is a label and you end up with a truncated data set.