search

ELEMKEP / bsc_lcs

Implementation of Brain Signal Classification via Learning Connectivity Structure
MIT License
16 stars 4 forks source link

Trouble shotting during reproducing experiments on the DREAMER dataset #2

Closed tczhangzhi closed 3 years ago

tczhangzhi commented 3 years ago

Hi, there! Thanks for your great work! I met some trouble during reading your paper and reproducing experiments on the DREAMER dataset.

https://github.com/ELEMKEP/bsc_lcs/blob/9972660a947d7a7244cd9eeab202857c87dd3112/data/lmdb_generate.py#L41

Could you please explain the process of generating Process.mat based on DREAMER.mat? It looks like you have concatenated the data of each TRAIL together. Have you performed any other operations? I would appreciate it if you could provide the code.

ELEMKEP commented 3 years ago

Thanks for your attention!

Due to the inconvenient interface of reading .mat file in python, the pre-processing for DREAMER.mat file was conducted on the MATLAB. The pre-processing in .mat file has nothing special but picking the necessary data from DREAMER.mat file, rearranging them, and cut the dataset by the first 60-second. I forgot to include the .m file in the repo, but it can be implemented in python, so I will implement it and update them.

It'll takes some time, and I'll send you e-mail with the .m file.

tczhangzhi commented 3 years ago

Thanks for your fast reply, it works well. And I believe most of the experimental results are stable after large amounts of epochs of training. Good job!