While trying to reproduce the results of your study, I had a hard time fining the downstream datasets you have used in your experiments and running the model on them. Below are described the issues for each dataset:
BCIC: There is already an existing open issue on the topic that the original version of the dataset is no longer available and the model cannot be ran on the current one.
ERN: Cannot be accessed after extensive search.
P300: Model throws a value error saying that "file does not start with a file id tag". I suspect a similar issue as with BCIC, that the version accessible online differs from the original one that you used to produce the results.
MMI: All model variations processed the dataset successfully. However, the obtained results slightly differ from the ones you have published. I was wondering, if there is a pre-or pos-processing step that you perform on the dataset or the obtained results and I am missing and, therefore, I cannot reach the same values. Are there other possible reasons for the mismatching results?
I wanted to ask if you could provide a way to access the original datasets that you used. This might help with all of the issues stated above.
While trying to reproduce the results of your study, I had a hard time fining the downstream datasets you have used in your experiments and running the model on them. Below are described the issues for each dataset:
I wanted to ask if you could provide a way to access the original datasets that you used. This might help with all of the issues stated above.